<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by ParallelDots on Medium]]></title>
        <description><![CDATA[Stories by ParallelDots on Medium]]></description>
        <link>https://medium.com/@ParallelDots?source=rss-62c5679fb60d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*VPbOVjuxjN7ThSqf.jpg</url>
            <title>Stories by ParallelDots on Medium</title>
            <link>https://medium.com/@ParallelDots?source=rss-62c5679fb60d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 05:02:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ParallelDots/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Visual Aesthetics: Judging a photo’s quality using AI techniques]]></title>
            <link>https://medium.com/@ParallelDots/visual-aesthetics-judging-a-photos-quality-using-ai-techniques-f0662667e5b2?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/f0662667e5b2</guid>
            <category><![CDATA[visual-analytics]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Tue, 09 Jan 2018 10:08:00 GMT</pubDate>
            <atom:updated>2018-01-09T10:08:00.510Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dsVZ5OTrJf6vsAP3.jpg" /></figure><p>Visual aesthetics has been shown to critically affect a variety of constructs such as perceived usability, satisfaction, and pleasure. However, visual aesthetics is also a subjective concept and therefore, presents its unique challenges in training a machine learning algorithm to learn such subjectiveness.</p><p>Given the importance of visual aesthetics in human-computer interaction, it is vital that machines adequately assess the concept of visual aesthetics. Machine learning, especially deep learning techniques have already shown great promise on tasks with well-defined goals such as identifying objects in images or translating from one language to another. However, quantification of image aesthetics has been one of the most persistent problems in image processing and computer vision.<br>We decided to build a deep learning system that can automatically analyze and score an image for aesthetic quality with high accuracy. Please check out our <a href="https://www.paralleldots.com/visual-analytics">demo</a> to check your photo’s aesthetic score.</p><h3>About the Research</h3><p>We came up with a novel Deep Convolutional Neural Network which can be trained to recognize an image’s aesthetic quality. We also came up with multiple hacks while training the algorithm to increase the accuracy.</p><p>In our paper published on <a href="https://arxiv.org/abs/1712.03382v1">arxiv</a>, we have proposed a new neural network architecture which can model the data efficiently by taking both low level and high-level features into account. It is a variant of DenseNets which has a skip connection at the end of every dense block. Besides this, we also propose training techniques that can increase the accuracy with which the algorithm trains. These methods are to train on LAB color space and to use similar images in a minibatch to train the algorithm, which we call coherent learning. Using these techniques, we get an accuracy of 78.7% of the AVA2 dataset. The state of the art accuracy on the AVA2 dataset is 85.6% which uses a deep Convolutional Neural Network with pretrained weights on the imagenet dataset. The best accuracy on the AVA2 dataset using handcrafted features is 68.55%. We also show that adding more data to our training set (from AVA dataset not included in AVA2) increases its accuracy to 81.48% on AVA2 Test Set, hence showing the model gets better with more data.</p><h3>Use-cases of Visual Aesthetics</h3><ol><li><strong>Selecting best profile photo for your social media account</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*a_cjmtR0oMKXekzv.jpeg" /></figure><p>App developers of social media sites can help their users decide which photo will suit best for their profile image. We all have faced anxiety while uploading photos on social media sites or changing our display pic. With our API integration, app developers can help their users look good, always!</p><ol><li><strong>Dating Apps</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/0*LmKVkx0Jkg31bEEH.jpg" /></figure><p>Ok, now this use-case may not appeal to the zen, non-materialistic folks among us but to be honest, dating leads to the most social anxiety. Dating landscape keeps changing as well and therefore, if you are active on dating apps, it’s important to choose your best photos to improve your chances for right swipes!<br>Dating App developers can easily integrate our APIs to help their users upload their best photos; the visual aesthetics model can also be fine-tuned if the developers want to optimise it on their data set.</p><ol><li><strong>AI enabled camera phones</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*LgE9Av5fiukJ7RDk.jpeg" /></figure><p>Recently Google has launched Pixel 2 and Pixel 2 XL which has a portrait mode. This phone offers the portrait mode even though it lacks the second lens that many other phones have. For example, the iPhone X, Galaxy Note 8, OnePlus 5… all these phones offer the portrait mode because they use data from two lenses. One lens captures the image, the other one captures the depth information, apart from providing some focal range magic for the blurred background. However, Pixel phone uses AI to give HDR+ images to users which are comparable to pictures clicked by a DSLR camera.</p><p>Similarly, mobile manufacturers can augment the capabilities of their native camera by integrating the visual aesthetic APIs to let their users know in real-time the quality of their photo even before taking a snap! This will enable your users to share their photos with confidence and you will end up creating a great differentiator for your brand at no additional hardware cost.</p><ol><li><strong>Virality in online content</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*pZ261ihkY9EQh3h6.jpeg" /></figure><p>Content is king, and it has become ever more difficult to write compelling content that resonates with your audience. However, the best content these days often have great images to complement them, and therefore, you’ve got to include something that will keep eyes moving down the page.<br>BuzzSumo did an analysis that covered over 1 million articles and found that the ones that had images every 75–100 words had more social shares. Using our visual aesthetics tool, you can quickly check how appealing your images are and accordingly, improve the virality of your blog post. You can check the demo <a href="https://www.paralleldots.com/visual-analytics">here</a>.</p><p>In this blog post, we have covered some of the use-cases of our visual aesthetics API. When machines become more competent than humans to judge such subjective content, it opens up a lot of possibilities to exploit which were not feasible yet. You can read more blogs on Visual Analytics <a href="http://blog.paralleldots.com/category/product/visual-analytics/">here</a>.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our text analysis<a href="https://www.paralleldots.com/text-analysis-apis"> APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f0662667e5b2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Sentiment Analysis: Concept, Analysis and Applications]]></title>
            <link>https://medium.com/@ParallelDots/sentiment-analysis-concept-analysis-and-applications-31344a37c981?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/31344a37c981</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[text-classification]]></category>
            <category><![CDATA[text-analytics]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Tue, 09 Jan 2018 10:05:06 GMT</pubDate>
            <atom:updated>2018-01-09T10:05:06.891Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*3zLqyxheWXqGvbK1.jpeg" /></figure><p><a href="https://www.paralleldots.com/sentiment-analysis">Sentiment analysis </a>is contexual mining of text which identifies and extracts subjective information in source material, and helping a business to understand the social sentiment of there brand, product or service while monitoring online conversations. However, analysis of social media streams is usually restricted to just basic sentiment analysis and count based metrics. This is akin to just scratching the surface and missing out on those high value insights that are waiting to be discovered. So what should a brand do to capture that low hanging fruit?</p><p>With the recent advances in deep learning, the ability of algorithms to analyse text has improved considerably. Creative use of advanced artificial intelligence techniques can be an effective tool for doing in-depth research. We believe it is important to classify incoming customer conversation about a brand based on following lines:</p><ol><li>Key aspects of a brand’s product and service that customers care about.</li><li>Users’ underlying intentions and reactions concerning those aspects.</li></ol><p>These basic concepts when used in combination, become a very important tool for analyzing millions of brand conversations with human level accuracy. In the post, we take the example of Uber and demonstrate how this works. Read On!</p><h3>Text Classifier — The basic building blocks</h3><p><strong>Sentiment Analysis<br></strong>Sentiment Analysis is the most common text classification tool that analyses an incoming message and tells whether the underlying sentiment is positive, negative our neutral. You can input a sentence of your choice and gauge the underlying sentiment by playing with the demo <a href="https://www.paralleldots.com/sentiment-analysis">here</a>.</p><p><strong>Intent Analysis<br></strong>Intent analysis steps up the game by analyzing the user’s intention behind a message and identifying whether it relates an opinion, news, marketing, complaint, suggestion, appreciation or query.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/666/0*YKSLCADWBznGXdFN.png" /><figcaption>Analyzing intent of textual data</figcaption></figure><p><strong>Contextual Semantic Search(CSS)</strong><br>Now this is where things get really interesting. To derive actionable insights, it is important to understand what aspect of the brand is a user discussing about. For example: Amazon would want to segregate messages that related to: late deliveries, billing issues, promotion related queries, product reviews etc. On the other hand, Starbucks would want to classify messages based on whether they relate to staff behavior, new coffee flavors, hygiene feedback, online orders, store name and location etc. But how can one do that?</p><p>We introduce an intelligent smart search algorithm called <strong>Contextual Semantic Search (a.k.a. CSS)</strong>. The way CSS works is that it takes thousands of messages and a concept (like <strong>Price</strong>) as input and filters all the messages that closely match with the given concept. The graphic shown below demonstrates how CSS represents a major improvement over existing methods used by the industry.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*9FS5XcTnJvrdIU3HUB-t7Q.png" /><figcaption>Existing approach vs Contextual Semantic Search</figcaption></figure><p>A conventional approach for filtering all <strong>Price </strong>related messages is to do a keyword search on <strong>Price </strong>and other closely related words like (pricing, charge, $, paid). This method however is not very effective as it is almost impossible to think of all the relevant keywords and their variants that represent a particular concept. CSS on the other hand just takes the name of the concept (<strong>Price</strong>) as input and filters all the contextually similar even where the obvious variants of the concept keyword are not mentioned.</p><p>For the curious people, we would like to give a glimpse of how this works. An AI technique is used to convert every word into a specific point in the hyperspace and the distance between these points is used to identify messages where the context is similar to the concept we are exploring. A visualization of how this looks under the hood can be seen below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/471/1*kj-0aV4KGv26unO2IgumOw.png" /><figcaption>Visualizing contextually related Tweets</figcaption></figure><p>Time to see CSS in action and how it works on comments related to Uber in the examples below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/676/1*Mm05Ob8NT5O7IjWvqRYFBg.png" /></figure><p>Similarly, have a look at this tweet:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1009/1*mncpMbFQYBi7QNDMeRCk_w.png" /></figure><p>In both the cases above, the algorithm classifies these messages as being contextually related to the concept called P<strong>rice</strong> even though the word <strong>Price</strong> is not mentioned in these messages.</p><h3>Uber: A deep dive analysis</h3><p>Uber, the highest valued start-up in the world, has been a pioneer in the sharing economy. Being operational in more than 500 cities worldwide and serving a gigantic user base, Uber gets a lot of feedback, suggestions, and complaints by users. Often, social media is the most preferred medium to register such issues. The huge amount of incoming data makes analyzing, categorizing, and generating insights challenging undertaking.</p><p>We analyzed the online conversations happening on digital media about a few product themes: <strong>Cancel, Payment, Price, Safety and Service.</strong></p><p>For a wide coverage of data sources, we took data from latest comments on Uber’s official Facebook page, Tweets mentioning Uber and latest news articles around Uber. Here’s a distribution of data points across all the channels:</p><ol><li>Facebook: <strong>34,173 </strong>Comments</li><li>Twitter: <strong>21,603 </strong>Tweets</li><li>News: <strong>4,245 </strong>Articles</li></ol><p>Analyzing sentiments of user conversations can give you an idea about overall brand perceptions. But, to dig deeper, it is important to further classify the data with the help of Contextual Semantic Search.</p><p>We ran the Contextual Semantic Search algorithm on the same dataset, taking the aforementioned categories in account (<strong>Cancel, Payment, Price, Safety, and Service</strong>).</p><h4>FACEBOOK</h4><p><strong>Sentiment Analysis</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*1IhRws4VYRQ4ptUqPEGzLQ.png" /><figcaption>Breakdown of Sentiment for Categories</figcaption></figure><p>Noticeably, comments related to all the categories have a negative sentiment majorly, bar one. The number of positive comments related to <strong>Price </strong>have outnumbered the negative ones. To dig deeper, we analyzed intent of these comments. Facebook being a social platform, the comments are crowded <em>random content, news shares, marketing and promotional content and spam/junk/unrelated content</em>. Have a look at the intent analysis on the Facebook comments:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/406/1*a-xKuw7k_29ZDSYIYr_rIw.png" /><figcaption>Intent analysis of Facebook comments</figcaption></figure><p>Intent analysis of Facebook comments</p><p>Thus, we removed all such irrelevant intent categories and reproduced the result:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*7kfWriHHUtlFIZl1K4EImw.png" /><figcaption>Filtered Sentiment Analysis</figcaption></figure><p>There is noticeable change in the sentiment attached to each category. Especially in <strong>Price</strong> related comments, where the number of positive comments has dropped from 46% to 29%.</p><p>This gives us a glimpse of how CSS can generate in-depth insights from digital media. A brand can thus analyze such Tweets and build upon the positive points from them or get feedback from the negative ones.</p><h4>TWITTER</h4><p><strong>Sentiment Analysis</strong></p><p>A similar analysis was done for crawled Tweets.<strong> </strong>In the initial analysis Payment and <strong>Safety</strong> related Tweets had a mixed sentiment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*2ES9VQlONzuZS7JEI0RChw.png" /><figcaption>Category wise sentiment analysis</figcaption></figure><p>To understand real user opinions, complaints and suggestions, we have to again filter the the unrelated Tweets(Spam, junk, marketing, news and random):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*9uyTsqKt7QalrD72nWI6RA.png" /><figcaption>Filtered sentiment</figcaption></figure><p>There is a remarkable reduction in number of positive <strong>Payment</strong> related Tweets. Also, there is a significant drop in the number of positive Tweets for the category <strong>Safety</strong>(and related keywords.)</p><p>Additionally,<strong> Cancel, Payment </strong>and <strong>Service</strong> (and related words) are the most talked about topics in the comments on Twitter. It seems that people talked most about drivers cancelling their ride and the cancellation fee charged to them. Have a look at this Tweet:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1010/1*zy9cJ6qeK2I2khqFJzaJGQ.png" /></figure><p>Brand like Uber can rely on such insights and act upon the most critical topics. For example, <strong>Service </strong>related Tweets carried the lowest percentage of positive Tweets and highest percentage of Negative ones. Uber can thus analyze such Tweets and act upon them to improve the service quality.</p><h4>NEWS</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*UZI9pGvlvr5yc9f4Aeg5Mw.png" /><figcaption>Sentiment Analysis for News headlines</figcaption></figure><p>Understandably so, Safety has been the most talked about topic in the news. Interestingly, news sentiment is positive overall and individually in each category as well.</p><p>We classified news based on their popularity score as well. The popularity score is attributed to the share count of the article on different social media channels. Here’s a list of top news articles:</p><ol><li><a href="https://www.nytimes.com/2017/02/02/technology/uber-ceo-travis-kalanick-trump-advisory-council.html?_r=0">Uber C.E.O. to Leave Trump Advisory Council After Criticism</a></li><li><a href="http://www.aljazeera.com/news/2017/01/deleteuber-users-angry-trump-muslim-ban-scrap-app-170129082003307.html">#DeleteUber: Users angry at Trump Muslim ban scrap app</a></li><li><a href="https://www.forbes.com/sites/alexkonrad/2017/04/07/uber-employees-hate-their-own-culture/#2df5598e6f95">Uber Employees Hate Their Own Corporate Culture, Too</a></li><li><a href="https://www.theguardian.com/commentisfree/2017/mar/03/uber-spreading-social-poison-travis-kalanick">Every time we take an Uber we’re spreading its social poison</a></li><li><a href="http://www.businessinsider.in/Furious-customers-are-deleting-the-Uber-app-after-drivers-went-to-JFK-airport-during-a-protest-and-strike/articleshow/56859124.cms">Furious customers are deleting the Uber app after drivers went to JFK airport during a protest and strike</a></li></ol><h3>Conclusion</h3><p>The age of getting meaningful insights from social media data has now arrived with the advance in technology. The Uber case study gives you a glimpse of the power of Contextual Semantic Search. It’s time for your organization to move beyond overall sentiment and count based metrics. Companies have been leveraging the power of data lately, but to get the deepest of the information, you have to leverage the power of AI, Deep learning and intelligent classifiers like Contextual Semantic Search and Sentiment Analysis. At <a href="http://www.karna.ai/">Karna</a>, you can contact us to license our technology or get a customized dashboard for generating meaningful insights from digital media. You can check the demo <a href="https://www.paralleldots.com/custom-classifier">here</a>.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our text analysis<a href="https://www.paralleldots.com/text-analysis-apis"> APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=31344a37c981" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Visual Analytics: Exploring #KendrickLamar on Instagram]]></title>
            <link>https://medium.com/@ParallelDots/visual-analytics-exploring-kendricklamar-on-instagram-ceb7aaa72034?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/ceb7aaa72034</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[visual-analytics]]></category>
            <category><![CDATA[instagram]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Thu, 28 Dec 2017 09:27:13 GMT</pubDate>
            <atom:updated>2017-12-28T09:27:13.171Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*v7A-wlQhpJ7IueiRPrXtwA.jpeg" /><figcaption>credits: Mic</figcaption></figure><p>In our previous blog <a href="http://blog.paralleldots.com/product/visual-analytics-instagrams-gopro-hashtag-ai/">post</a> of Visual Analytics we discussed Instagram’s #gopro hashtag using AI. In this blog post, we will be exploring #Kendrick Lamar on Instagram. Kendrick Lamar’s DAMN released this April with tons of rave reviews. Christopher R. Weingarten of <em>Rolling Stone</em> writer describing it as a combination of <a href="http://www.rollingstone.com/music/albumreviews/review-kendrick-lamar-damn-album-w477376">“the old school and the next-level.”</a> It marked his third #1 album on the <em>Billboard</em> 200 chart, and the single “Humble” became his first #1 as a lead artist on the <em>Billboard</em> Hot 100. We explored Instagram posts mentioning the hashtag #KendrickLamar to deep dive into what the world has to say for Kendrick Lamar using Visual Analytics.</p><p>Our AI-powered Visual Analytics capabilities can extract visual information such as Gender, Emotion, Key Tags present in the picture. Have a look at the analysis of this picture uploaded by a user on Instagram at Rolling Loud Festival 2017 which was headlined by Kendrick Lamar:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/473/1*wIeiycpeqD7jb3VhoPrlZg.png" /></figure><p><strong>Username:</strong> anisecrane<br><strong>Caption:</strong> Incredible weekend, incredible company #rollingloud2017 #kendricklamar #asaprocky<br><strong>Likes:</strong> 246<br><strong>Comments: </strong>4<br><strong>Location: </strong>Bayfront Park<br><strong>Gender and Emotion: </strong>[Man, Happy],[Woman, Happy]*<br><strong>Tags: </strong>[Sunglasses, Pajama, Sarog, Feather_boa]</p><p>We ran this analysis on 13,992 images mentioning the hashtag #KendrickLamar. Let’s dive right into the collective insights we got from them.</p><h4>DEMOGRAPHICS</h4><p>The hashtag #KendrickLamar was populated by male fans as almost a three-fourth of total humans present were men.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/628/1*4DBfAFyTpJZ0ux4Avqx6Fg.png" /><figcaption>Gender Breakdown</figcaption></figure><p>Emotion Detection</p><p>Our AI algorithms can detect the emotion expressed by the people in the photos. Here’s a breakdown of emotions detected in the images crawled.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/686/1*aMY8qpA1F0OWdCz3f0TqEQ.png" /></figure><p>Very few people expressed negative emotions such as sadness and anger. More than a third of the emotions expressed was happiness.</p><h4>TAGS</h4><p>Here’s a distribution of top 10 most frequently detected tags in the photo. You can see relevant tags such as Microphone, Monitor, Spotlight and Stage along with some generic ones such as Jeans, Jersey and Sunglasses.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*KwDkAkTgJPmQhweVL8MNCA.png" /></figure><p><strong>Top Locations</strong></p><p>The top locations were all cities in the United States with Los Angeles occurring the most with 883 mentions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*G0I3f78W46h77Ys1mRsnkg.png" /></figure><p>Such an analysis can help brands, marketers and researchers to understand what is being shared about their hashtags on visual social media without going through the pain of analyzing each image.</p><p>Visual analytics opens up a plethora of analytics opportunities ranging from market research stage for a product launch to analyze the performance of promotional campaigns. You can check the demo <a href="https://www.paralleldots.com/visual-analytics">here</a>.</p><p>At <a href="http://www.karna.ai/">Karna</a>, we do customized in-depth text and visual analytics research for multiple brands. You can reach to us <a href="http://www.karna.ai/">here</a> for any similar analysis for your brand.</p><p>The above study was carried out by <a href="http://karna.ai/">Karna AI</a>, Market Research division of ParallelDots Inc.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ceb7aaa72034" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Automated Caries Detection on Bitewing Radiographs Using Deep CNNs]]></title>
            <link>https://medium.com/@ParallelDots/automated-caries-detection-on-bitewing-radiographs-using-deep-cnns-d2b4a68c64ab?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2b4a68c64ab</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[healthcare]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Fri, 01 Dec 2017 12:12:15 GMT</pubDate>
            <atom:updated>2017-12-01T12:12:15.841Z</atom:updated>
            <content:encoded><![CDATA[<p>Research has established that a large percentage of dental caries escape identification in routine dental examinations, even when such examinations include dental x-rays. Certain types of caries, like occlusal caries, appear to be easier to find through a normal clinical examination or x-ray review, whereas diagnosing other types of caries, such as caries below the surface of the tooth, interproximal caries, and root caries, is often not as reliable.</p><p>At <a href="https://paralleldots.xyz/">ParallelDots, Inc</a>. we took the challenge to find these dental anomalies with human-level accuracy and build a reliable diagnostic tool for the dentists. In this blog post, we will discuss the recent study we did for our automated caries detection system, putting our system against three practicing dentists from North American clinics. We found that our system had a higher agreement (F Score) with clinically verified ground truth than all three individually (the difference between system’s F Score and average F-Score of the dentists is over 17%). Our system has higher sensitivity with respect to Dentists individually and hence can be used as a tool to ease the work of dentists by suggesting them possible caries they can then verify and treat. A breakdown of metrics in our test is given in the table below. Please note that the metrics Recall and Precision are names Machine Learning community generally uses for Sensitivity and Positive Predictive Value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/929/0*BbHiyjTw4fWk3aT-.png" /></figure><p>Our <a href="https://arxiv.org/abs/1711.07312v1">paper</a> detailing the experiments performed has been accepted at the NIPS 2017 Workshop on Machine Learning for Health being held on the theme “What Parts of Healthcare are Ripe for Disruption by Machine Learning Right Now?”. NIPS (Neural Information Processing Systems) is among the topmost Machine Learning conferences globally and has two tracks of papers and multiple focused workshops.</p><h3>How does the automated caries detection technology work?</h3><p>The first step is to gather a dataset to train AI algorithm on. We obtained an IRB approval and collected data from multiple clinics across the US, where dentists marked clinically verified cavities on their XRay archives. 3000 radiographs with cavities annotated were collected as a part of this exercise. The data is then anonymized as soon as the clinician uploads it to our HIPAA compliant server and the dataset thus is ready to be used by our Data Science team.</p><p>The next step for the Data Science team is to decide:<br>a. how to model this as a machine learning problem<br>b. How to calculate the accuracy of the machine learning system thus created.<br>Generally, the processes a and b go hand in hand (for example you would measure accuracy for something you would model as a pixelwise classification task using dice coefficient, for an object detection task as Intersection over Union or for a classification task as accuracy). In this case, however, the problem arises due to the following complications:</p><ol><li>Dental Cavities are irregular shaped, thus there is no way to mark them exactly by the dentists. At best we can get approximate polygon annotations for dental caries. Although modeling the cavity detection problem as Dense Classification task is pretty obvious, a dice score will not be a fair measure of accuracy on such approximate annotations.</li><li>While average IoU score might be a good measure it is hard to visualize how effective the model is just by this number in the real world. Hence we put a more practical criterion for the model to be evaluated. We evaluate the model as if it were a search engine to find cavities. A search engine is evaluated based on measures of precision, recall and F Score. Precision denotes the number of correct searches among total results returned by the cavity search. Out of the total cavities present, the number of cavities that the system can detect is called Recall.</li></ol><p>A radiograph containing caries and the output processed by our algorithm is shown below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/914/0*-ajsZFll3x4NFqt6.png" /></figure><p>With this research, we are aiming to assist dentists in analyzing and detecting dental problems. The automated caries detection system can provide a solid assistance to dentists and make their work easier, faster and more accurate so that they can effectively concentrate on rectifying the dental problems much faster.</p><p>At ParallalDots, we are working on challenging problems in Healthcare, NLP and Image Classification. Please watch this space for more updates on the research work at ParallelDots.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our text analysis<a href="https://www.paralleldots.com/text-analysis-apis"> APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2b4a68c64ab" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Text Analysis in Excel: Real world use-cases]]></title>
            <link>https://medium.com/@ParallelDots/text-analysis-in-excel-real-world-use-cases-3ee0a0ea1ba9?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/3ee0a0ea1ba9</guid>
            <category><![CDATA[text-analytics]]></category>
            <category><![CDATA[excel]]></category>
            <category><![CDATA[sentiment-analysis]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[nlp]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Wed, 29 Nov 2017 07:21:32 GMT</pubDate>
            <atom:updated>2017-11-29T07:21:32.962Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vU52DS3ECG9ILxC1.jpg" /></figure><p>Last week, we launched <a href="https://www.paralleldots.com/excel-plugin">ParallelDots Excel add-in</a>, a solution for using ParallelDots NLP APIs to do text analysis on unstructured data without writing a single line of code. The Excel add-in is very easy to use and provides a convenient, yet effective solution for your text analysis needs. In an earlier <a href="http://blog.paralleldots.com/product/announcing-excel-add-paralleldots-ai-apis/">post</a>, we provided you with detailed information of how the excel add-in works. In this post, we will discuss some real-world use cases where you can use the Excel Add-in to raise your analytics game without spending a fortune on building a data science team.</p><h3>Text analysis on product reviews from E-commerce sites, Facebook Pages and other review sites</h3><p>You can analyze a corpus of customer reviews to understand the general impression about your product. The Excel add-in works on ParallelDots AI APIs, which are being used extensively by developers and enterprises to empower their analytics since the last two years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/796/0*3-kzYzSmQuiwZBRC.png" /></figure><p>Furthermore, you can run keyword analysis on positive and negative sentiment sentences to discover why people like or dislike your product. Such an analysis will enable you to understand the key phrases contributing to the overall sentiment around your product. For example, a phone manufacturer company can analyze reviews from e-commerce sites, social media, and tech review blogs for their models. They can then extract keywords for both positive and negative sentiment sentences to find which features (screen, battery life, camera, storage, etc.) users liked or disliked about a particular phone model.</p><p>Additionally, you can go a level deeper by analyzing the intent of the product reviews and categorize them to identify whether the review counts as the feedback, opinion, query or spam. This will help in filtering essential reviews and replying/acting on them faster.</p><p>The analyzed data, i.e. sentiment, keywords and intent, can be correlated with your internal business metrics such as marketing spends, sales data etc. to find actionable insights.</p><h3>Analyse open-ended user feedback from surveys</h3><p>Open-ended feedback forms have proven to be more useful than close-ended ones(Folks at our market research division, Karna AI, have written a extensive blog on analyzing open-ended user surveys. You can read it <a href="https://blog.karna.ai/challenges-with-analyzing-open-ended-text-are-disappearing-a348d55e2ab2">here</a>) but analyzing them can be very time-consuming and expensive task.</p><p>Our Excel Add-in is a lifesaver when it comes to analyzing open-ended surveys. Once you have exported all the survey responses in xlsx or csv format and installed the add-in, you can start analyzing the responses straight away without writing a single line of code. The Excel add-in makes it simple for anyone and everyone to perform AI-powered text analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*hWgGxxT389faNnJU.png" /></figure><p>Methodology can be same as the one mentioned in the product reviews analysis, and you can incorporate entity extraction function along with keyword analysis to discover broader entities (product name, person name etc.) talked about in the surveys.</p><h3>Text Analysis for financial content</h3><p>Financial data such as earnings call, news articles, transcripts etc. contain lot of rich information which can form basis for your fundamental investment and financial research. The BFSI sector is heavily equipped with textual data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Luo3FbqVCR34rwgh.jpeg" /></figure><p>Since excel remains an integral part of any financial analyst’s toolkit, ParallelDots Excel add-in can help to correlate many of the structured data points found in financial documents (balance sheets, trading history, P&amp;L statements) with those found in unstructured financial documents.</p><h3>Analyzing social media content from your social listening tool</h3><p>While social listening tools do a good job of tracking every conversation about your brand or product on the internet, building good text analytics capability is only considered as an afterthought in most of these tools. The data from such social listening tools analyzed by a powerful text analytics tool like ParallelDots Excel add-in can provide a good overview of your social presence.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iLhKvmbFnYF3gl4m.jpeg" /></figure><p>Further, listening tools that do provide good text analytics capabilities can be an expensive tool to use when compared to the Excel add-in like ParallelDots.</p><p>These listening tools give a very convenient option to export all the conversations in csv format. You can then analyze these conversations similar to product reviews to get actionable insights such as sentiment attached to the conversation, keywords and key-phrases mentioned, emotions detected in the conversation, entities mentioned, etc.</p><h3>Empowering business analysts to carry out text analysis on unstructured datasets</h3><p>Business analysts capture, analyze and document business needs and changes. A business analyst is usually equipped with huge textual corpus from different domains such as market research, financial data, historical viewpoints, etc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5joeiaVn_aQojy1d.jpeg" /></figure><p>ParallelDots Excel Add-in can help analysts up their game since excel is heavily used in analyzing the data points in this sector. Business analysts can process the information faster using multiple functions from the ParallelDots Excel Add-in. They can gather relevant insights such as extracting the entities, categorizing products, services or markets based on sentiment and classifying text based on the IAB categories etc.</p><p>The use cases of ParallelDots Excel add-in are not limited to the ones mentioned above. Essentially, you can use the add-in whenever you require to process a large corpus of unstructured textual data. It is super easy to use, just <a href="https://www.paralleldots.com/excel-plugin">download</a> the add-in, <a href="https://www.paralleldots.com/sign-up">sign up</a> for your unique API key and start analysing your content. Please check out the <a href="https://www.paralleldots.com/excel-docs">documentation</a> to know more about the different functions available in the add-in.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our text analysis<a href="https://www.paralleldots.com/text-analysis-apis"> APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3ee0a0ea1ba9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Visual Analytics of Instagram’s #gopro hashtag with AI]]></title>
            <link>https://medium.com/@ParallelDots/visual-analytics-of-instagrams-gopro-hashtag-with-ai-1984f809797e?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/1984f809797e</guid>
            <category><![CDATA[visual-analytics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[analytics]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Sat, 25 Nov 2017 12:14:36 GMT</pubDate>
            <atom:updated>2017-11-25T12:14:36.070Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/640/0*7byqBgQ8AE-KEWCz.jpeg" /></figure><p>Images have become a very common medium of human expression on the internet with the coming up of social networks. Facebook is the biggest repository of digital images ever. This trend is only going to intensify given the emergence of image first platforms like Instagram and Snapchat, also called “Visual Social Media”. Marketers and analysts generally find it really hard to gather insights from these visual social media, because its hard to quantify images being shared. The answer lies in automated image analytics i.e. Visual Analytics that can process the visual information and derive conclusions. Until very recently, it wasn’t even possible to do such automated analytics but times have changed. Now <a href="https://www.paralleldots.com/visual-analytics">Visual Analytics</a> tools are used to get the analysis of images.<br>With the recent success of deep learning models called convolutional neural networks in automated perception tasks, AI has matured enough to act as a proxy to human observers to document what is being shared. These AI algorithms can actually make sense of what the content of the image is (for instance, it can see that the image contains dogs or cats or an apple) and even asses various quantities for us (Microsoft recently released a demo where they could estimate the age of the person in the image). Extracting such information from the images automates the categorization and processing of visual data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4hMVFdJLT0WpvZAY.png" /></figure><p>When you put an image for input, the AI algorithm returns with the visual information associated with that image. The output, for example can tell what all objects are present in the image , how many humans in the image along with their age and emotions, if the image is taken outside or indoors and even the relationship among the objects in the image (for instance, it can recognize a human and a frisbee in the image and can also recognize that the human is throwing a frisbee and not the other way around). We decided to run these AI algorithms and check out what we could deduce. While we continue working on our own versions of these algorithms, here we used our own <a href="https://www.paralleldots.com/visual-analytics">Visual Analytics APIs</a> on top of social media images which could return the image description, tags and emotions of the input image. Take a look at the following image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/0*l6VFPkmNJ1Bx-UXS.jpeg" /></figure><p>For this image, the <a href="https://www.paralleldots.com/visual-analytics">visual analytics</a> returns with the these results.</p><p><em>Emotion: happiness</em></p><p><em>Tags: [“indoor”, “table”, “cake”, “sitting”, “top”, “decorated”, “birthday”, “pink”, “girl”, “food”, “plate”, “topped”, “small”, “front”, “little”, “woman”, “bed”, “desk”, “holding”, “white”, “young”, “large”, “standing”, “bear”, “room”, “phone”]</em></p><p>Talking of images, Instagram is probably the largest social media platform dedicated entirely for featuring images tagged with one or the other hashtags. <a href="https://www.instagram.com/explore/tags/gopro/?hl=en">#gopro</a> is an evergreen hashtag on Instagram almost always. Go pro is known for its action cameras and mobile apps. We crawled all the public images using this hashtag uploaded in last seven days. Upon the image analysis of a total of 727 images using the technique specified above and aggregating results, we came up with the following insights. (Please note there would be way more than 727 private images, which we did not work on)</p><p>The most common emotion was happiness that was exhibited in 531 images. This was followed by the no emotion(187), surprise (4), anger (3) and sadness (1).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/605/0*_LyearVxU9rb7zeZ.png" /></figure><p>For the next insight, we have compiled a graph of the top 10 tags with their frequencies that were found in the photo using <a href="https://www.paralleldots.com/visual-analytics">visual analytics</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/605/0*NGi-bhA1Ef3fTn89.png" /></figure><p>As you might be able to decipher, they are more generic image tags, showing demographical and indoor/outdoor distribution of images. Also white is the most popular color.</p><p>The remaining tags are arranged in the following word cloud (the size of the word corroborates with the frequency of tag in the images).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*otHH2yud0OaD4NKy.png" /></figure><p>Judging by the tags we have assembled, it is quite evident that most of the images are taken outdoors either in the vicinity of natural offerings like ocean, grass, rivers, mountains or doing outdoor sporty activities. Well, we don’t blame them as the best pictures are clicked when you have a natural light in sight.</p><p>In general, such an analysis can help brands, marketers and researchers to understand what is being shared about their hashtags on visual social media without going through the pain of analyzing each image.</p><p>please write to us at apis@paralleldots.com in case of any queries or feedback.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our <a href="https://www.paralleldots.com/visual-analytics">Visual analytics APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1984f809797e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Announcing Excel Add-in for ParallelDots AI APIs]]></title>
            <link>https://medium.com/@ParallelDots/announcing-excel-add-in-for-paralleldots-ai-apis-8e5c61ec4387?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/8e5c61ec4387</guid>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Fri, 17 Nov 2017 09:01:22 GMT</pubDate>
            <atom:updated>2017-11-17T09:01:22.364Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/620/0*5NBOggfX3oNfhnr7.png" /></figure><p>Our Excel Add-in provides state-of-the-art text analysis capabilities without writing a single line of code. The add-in also comes in handy when you need to run text analysis in batches on a large corpus of text and discover insights in them (such as user-generated content from social media campaigns, earnings call transcripts, open-ended user feedback etc.). You can export all your data from any BI tool you use in xlsx (or CSV format) and install our plugin to annotate the data with sentiment, emotions, intent etc. and analyze them from the comfort of your spreadsheet.</p><p>In this post, we will show you how to use our Excel Add-in to add an important tool to your text analytics and text mining capabilities.</p><h3>Installing and using the Add-in</h3><p>Before starting with the Excel add-in you need to <a href="https://www.paralleldots.com/">sign up</a> for a free ParallelDots AI APIs account. Upon filling the form on our website and verifying your email, you will receive an API key in your inbox. Please store the API key in a secure location as unauthorized access to your API keys will result in early exhaustion of your quota. In case you feel you have lost your API key or someone else might have access to it, please drop an email to <a href="/service/mailto:support@paralleldots.com">support@paralleldots.com</a> and we will immediately invalidate your old API key and generate a new one for you.</p><p>Upon receiving the API key, <a href="https://www.paralleldots.com/excel-plugin">download</a> the Excel add-in and follow the on-screen instructions to install the add-in. Upon successful installation of the add-in, your excel screen should have a ParallelDots button in the top navigation bar as shown below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*6xYVqIIkxhSDmawU.png" /></figure><p>Upon clicking the Show/Hide Pane, you can toggle the settings screen to activate your add-in account. Enter the API key generated in the previous step in the Task Pane, and click submit to start analyzing your data.</p><p>That’s it, now you can bring any type of unstructured textual content and analyze with our add-in. In the next section, we will explain the different functions available in the add-in to perform different types of text analytics.</p><h3>Text Analytics Functions</h3><p>Our excel add-in lets you access the following APIs:</p><p><strong>Sentiment Analysis</strong></p><p>Sentiment Analysis using Excel can be accessed by invoking function<strong>paralledots_sentiment</strong> as shown in the screenshot below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/877/0*aSL6xxfMIs9O3yZc.png" /></figure><p>In case, you need to know the probability of sentiment label returned by <strong>paralleldots_sentiment</strong> function above, you can do so by invoking function <strong>paralleldots_sentiment_&lt;sentiment_label&gt;_probability</strong> where sentiment_label can be neutral, positive or negative.</p><p>Details of Sentiment Analysis API can be found <a href="https://www.paralleldots.com/sentiment-analysis">here</a>.</p><p><strong>Multilingual Sentiment Analysis</strong></p><p>Our Sentiment Analysis algorithm also supports three other languages apart from English, namely Spanish, Portuguese and Chinese. Analysis for these languages can be invoked by calling the <strong>paralleldots_multilang_es, paralleldots_multilang_pt and paralleldots_multilang_cn</strong> functions respectively. Their confidence scores can be also be retrieved by appending “_confidence” at the end of each language’s function as shown in the screenshot below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/787/0*aZEGPAjY3616_Jh2.png" /></figure><p><strong>Keywords Extraction</strong></p><p>Can be invoked by calling <strong>paralleldots_keywords</strong> function. Returns a list of important keywords in the input text. Details of this function can be found <a href="https://www.paralleldots.com/keyword-generator">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/876/0*SND8eWxHO4W9RxR1.png" /></figure><p>Text Classification</p><p>Can be invoked by calling <strong>paralleldots_taxonomy</strong> function. Returns a list of probable categories that the text can belong to. More details on this function can be found <a href="https://www.paralleldots.com/text-classification">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/858/0*lhnGe6HvFbbSkAef.png" /></figure><p>Entity Extraction</p><p>Can be invoked by calling <strong>paralleldots_ner_organization, paralleldots_ner_person </strong>or<strong> paralleldots_ner_place</strong> to extract organization name or person or place respectively.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/862/0*yFOGq5T439rKF8Sj.png" /></figure><p>More details on this function are available <a href="https://www.paralleldots.com/named-entity-recognition">here</a>.</p><p>Semantic Analysis</p><p>Can be invoked by calling <strong>paralleldots_similarity</strong>. The output is a score between 0–5 where score close to 5 indicates the two sentences are very close in meaning (semantically similar) while scores closer to 0 indicate the two sentences are very disparate in meaning.</p><p>Note that in order to use this function, you need to run it on two sentences, more details and use-cases can be found <a href="https://www.paralleldots.com/semantic-analysis">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/871/0*UdR3b5vVo4vbJ6hV.png" /></figure><p>Emotion Analysis</p><p>Can be invoked by calling paralleldots_emotion to get the emotion label and <strong>paralleldots_emotion_&lt;label&gt;_probability </strong>to find the probability score for different emotions as shown in the screenshot below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/698/0*e4C0dAOeJHY3CQ4B.png" /></figure><p>More details on emotion function is available <a href="https://www.paralleldots.com/emotion-detection">here</a>.</p><p>Intent Analysis</p><p>Can be invoked by calling paralleldots_intent to find the intent of the input text and paralleldots_intent_&lt;label&gt;_probability to find the probability scores of the different intent labels available.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/0*ADC4dYe5Ge9RzKBZ.png" /></figure><p>More details on Intent function is available <a href="https://www.paralleldots.com/intent-analysis">here</a>.</p><p>Abusive Content Classifier</p><p>Can be invoked by calling<strong> paralledots_abuse</strong> to classify input texts as abusive or not. Confidence score (probability score) can be found by calling <strong>paralleldots_abuse_confidence</strong>.</p><p>More details are available <a href="https://www.paralleldots.com/abusive-content">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/738/0*I-SmqaRWe4gQFE22.png" /></figure><p><strong>Security and Privacy</strong></p><p>Excel add-in is built on our APIs which means that your data is processed on our servers to get the final output. We take user privacy very seriously at ParallelDots and our privacy policy can be accessed <a href="https://www.paralleldots.com/terms-and-conditions">here</a>. All the user data is stored according to our privacy policy ensuring high standards of security.</p><p>However, in some cases due to contractual obligations or otherwise, the user may want to keep the data in-house in which case we can deploy these algorithms on-premise and build the plugin accordingly. Please <a href="https://www.paralleldots.com/contact-us">send us a request </a>to deploy these APIs on-premise and any custom function that you want us to build.</p><p>Excel add-in can be a powerful and cost-effective way to start adding cutting-edge intelligence to your analytics. Our excel add-in lets you analyze 100 sentences for each function every day for free without giving your credit card details. We hope this will allow you to understand the different functions provided in the add-in and understand the use-case before upgrading to one of our paid plans for large-scale usage.</p><p>Please write to us at apis<a href="/service/mailto:support@paralleldots.com">@paralleldots.com</a> in case of any queries or feedback.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our <a href="https://www.paralleldots.com/text-analysis-apis">text analysis APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8e5c61ec4387" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data Tagging in Medical Imaging — Diving Deep into the Processes]]></title>
            <link>https://medium.com/@ParallelDots/data-tagging-in-medical-imaging-diving-deep-into-the-processes-ac45eef92217?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/ac45eef92217</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[google-cloud-platform]]></category>
            <category><![CDATA[medical-imaging]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Mon, 13 Nov 2017 10:48:16 GMT</pubDate>
            <atom:updated>2017-11-13T10:48:16.950Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Note: This is the second post of the series Data Tagging in Medical Imaging. You can find the first post </em><a href="http://blog.paralleldots.com/healthcare/data-tagging-medical-imaging-overview/"><em>here</em></a><em>.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*rZcb7TfuMqpyG8Mi.png" /></figure><p>In the previous post of the series, Data Tagging in Medical Imaging, we gave you an overview of the kind of processes that you must put in practice to scale your data tagging engine. In this blog we will thoroughly discuss how to come up with these processes and things to consider before finalizing and formulating these processes. We will be discussing what these processes are and how they affect the data tagging process.</p><p>First of all, you need some resources to set up these processes. To set up them up, you need to make sure the availability of the following:</p><ul><li>A team of medical professionals that can annotate the data</li><li>Data annotation software where the medical professionals will annotate the data</li><li>And obviously…DATA!</li></ul><p>So, let’s deep dive into these processes; what are they and what is their importance and role in efficient data tagging. We have listed down three main processes. Ideally, this is all one large workflow process designed to help you annotate your medical dataset in a faster, scalable and accurate way.</p><ul><li>Data Flow</li><li>Quality Assurance/Quality Control</li><li>Providing the data to the data scientists</li></ul><p>Let’s discuss these aforementioned processes in details below:</p><p>Data Flow</p><p>It is really critical to define how the data will flow through all the stakeholders at different stages.</p><ul><li><strong>Data Storage:</strong><br>As data requirements for building a real-world AI solution is huge, you need to know where will be storing terabytes of data. You can either store it on the cloud using a number of cold storage such as Google Coldline, Google Nearline and Amazon Glacier etc. or choose to store it locally — <em>however, until you have a strong scalable local infrastructure with perfect power backup, we recommend that you use cloud for all the data flow. Cloud just gives you a lot of flexibility — and the storage is not that costly. </em>Other factors should be considered as well while deciding your storage facility such as the quantity of data you are storing and how many times will you retrieve the data from your storage etc.<em> </em>Noticeably, cold storage is used to store inactive data and hence, they are extremely cost-effective. But the data retrieval is costlier with cold storage platforms. So, it is advised to keep some amount of training data locally, using devices such as NAS(Network Attached Storage), which makes retrieval of a small amount of data easier.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/770/0*Q7JFZZvYRzU4nwDw.png" /></figure><ul><li><strong>Serving the Data:</strong><br>Medical professionals annotate the data that is served to them. This data is served within the data annotation software developed by your software development team. All the medical professionals who are going to annotate the data should have proper login credentials for better visibility and progress tracking. There are many caveats to consider while serving the data to the medical professionals based on their experience. In our data tagging engine, a senior radiologist will have different privileges than a junior radiologist, based on the roles defined. Similarly, you also have to ensure that all stakeholders see the information they need to see within the software. For example, non-personal information on the subject such as age bracket, gender, morbidity like diabetes etc. This supplies additional information to the medical professionals. Thus, allowing them to annotate accurately.</li></ul><p>Quality Assurance/Quality Control</p><p>This is perhaps the most important process in your workflow. The entire purpose of this exercise is to create a dataset that could be used to train the machines. The models that you create are only as accurate as the training dataset. Following are the things that you could do to create a strong QA/QC layer in your workflow.</p><ul><li><strong>Verification of annotations:<br></strong>Each and every tag needs to be thoroughly verified by a medical professional with a different privilege level. This double verification ensures a lesser error rate in the annotated dataset.</li><li><strong>Feedback to medical professionals:</strong><br>It is essential to continuously monitor the performance of the medical professionals and subsequently provide them the feedback. It could happen that some medical professionals are very good at annotating NCCT Head but not MRI Head. These things need to be learned and then be acted upon accordingly.</li></ul><p>Providing annotated data to the Data Scientists:</p><p>This is final layer — and there is a lot of scope for miscommunication in this layer. The entire purpose of annotating all the medical data is to provide it to the Data Science team for them to build the model. The Data Science team is always running a number of experiments for which they need a varied amount of data. In between all the experiments, it is very easy for the Data Scientists to be lost with a stream of annotated data coming from the Tagging Engine. This makes maintenance of data flow logs really important.</p><p>In this blog, we discussed the initial requirements of setting up a scalable tagging engine and core processes involved. Processes such as Data flow, QA/QC and training of medical professionals is critical in deploying a smart tagging engine. In achieving this, the proper formulation of the aforementioned processes is critical.</p><p>In the next blog of this series, we will discuss the most important aspect of the data tagging capabilities, Quality Control. You can find the first blog of the series <a href="http://blog.paralleldots.com/healthcare/data-tagging-medical-imaging-overview/">here</a>. Please watch this space for more.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our <a href="https://www.paralleldots.com/text-analysis-apis">text analysis APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ac45eef92217" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data Tagging in Medical Imaging — An overview]]></title>
            <link>https://medium.com/@ParallelDots/data-tagging-in-medical-imaging-an-overview-d57907c44ae6?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/d57907c44ae6</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[artificial]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Sat, 11 Nov 2017 10:35:56 GMT</pubDate>
            <atom:updated>2017-11-11T10:36:10.314Z</atom:updated>
            <content:encoded><![CDATA[<p>Today, as we are witnessing the era of smart AI-driven solutions which are empowering humans to automate tasks that require a certain level of cognition. One major reason for this shift in developing AI-driven products is the availability of a large amount of data. As we, humans, tend to learn through various experiences throughout our life, machines learn and automate tasks based on the data fed to them.</p><p>From our experience in developing vertical agnostic AI-first products, we are well aware of the importance of the availability of quality data and subsequently developing a smart data tagging process. In this series of blog posts, we’re going to talk about the importance of data tagging in medical imaging, where we are developing computer vision technologies to better assist doctors. This particular piece gives an overview of all the important things to consider when you are tagging the medical data for an AI-powered algorithm to use.</p><p>Important aspects of Data Tagging</p><p>Before you even begin working on tagging medical images, it is important to plan different aspects beforehand. We are talking about handling terabytes of medical imaging data. All the processes, workflows and other specs need to be finalized before you begin tagging the data. Let us walk you through the aspects which we should take in consideration for developing a smart data tagging.</p><h4>Deciding what to tag; and how to tag?</h4><p>This is the most important task — you should know what you want to tag. Considerable efforts need to be put in to understand the data you want to capture through the tagging. Determining the exact list of tags, levels of tagging and properties of each of these tags is extremely critical and, if done properly, can save you a lot of money and time. For example, to build a dataset of subtypes of hemorrhage, one needs to make a list of all types of intracranial hemorrhage such as subarachnoid hemorrhage and subdural hemorrhage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/547/0*caS1Qi-hym12nS-Z.png" /><figcaption><em>Tagging subarachanoid hemorrhage affected parts</em></figcaption></figure><p>Additionally, it is important to understand each tag in your list carefully as it comes with properties of its own. The tagging engine should be broad enough to include the properties of each tag. For example, intraparenchymal hemorrhage can be easily tagged by making an outline, whereas a pathology like cerebral atrophy is nearly impossible to tag by making a mere outline. Such an annotation would be a property of a slice in the entire CT scan. This exercise has to be done for the entire list so as to maintain the quality of tagging.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/520/0*VDtmqd-nO_goAYzw.jpg" /><figcaption><em>Difficult to mark cerebral atrophy using outlines. Credits: </em><a href="http://www.startradiology.com/"><em>StartRadiology</em></a></figcaption></figure><h4>Training the professionals:</h4><p>Processes: It is important to define all the processes right in the beginning and subsequently make sure that all the stakeholders understand them. These processes will enable you to build a faster, scalable and accurate tagging engine. Each aspect such as data flow, quality control and training processes is a broader topic of discussion, we will be discussing them in detail in the next few blogs.</p><p>Handshakes: A perfect data flow and synchronization have to be there in the involved parties; the Data Science team, Medical Professionals and the Software Development team.</p><ul><li>Data Science team: They have a very critical role to play in the engine. At the end, the output of the tagging engine is meant for them to train the algorithms. They need to explain the software development team the format in which they need the data. The team also helps in figuring out which tag should be annotated in which way.</li><li>Medical Professionals: They should be provided with all the required information which enables them to tag as accurately as possible. For example, if you want to tag ‘bone fracture’ in NCCT head, bone window is a must for them. Specifications of the software on which they will annotate the data should be planned after taking into account all the requirement of the medical professionals who will use the software.</li><li>Software Development team: They should know all the processes involved as well as the requirement from both Data Science team and the Medical Professionals so as to make sure the tagging engine has been developing according to their needs.</li></ul><p>Quality Control: Ensuring high quality of the tagged data is most important aspect of data tagging process. The AI you will train will only be as accurate as the data that it is fed with. Many levels of quality control measures should be in place such as double verification of each tag, privilege level of the medical professionals etc. Subsequently, the software design and development need to align with all the Quality Control measures.</p><p>Technology: Developing a cloud-based user-friendly tagging engine is necessary for faster and scalable operations. It also ensures easy coordination among all the stakeholders. Creating a tagging engine is complex, since it has various stakeholders with different requirements altogether and requires processing of terabytes of data. Hence, it requires a well thought of database and storage architecture.</p><p>At ParallelDots, we have built a scalable, accurate, and faster tagging engine. For developing products powered by AI, fine control over data acquisition, processing and deployment is absolutely necessary. In this post, we gave you an overview of the important aspects of data tagging capacity. In our next few posts of this series, we will discuss each of them in detail. Watch this space for more.</p><p><a href="https://www.paralleldots.com/">ParallelDots AI APIs</a>, is a Deep Learning powered web service by <a href="https://paralleldots.xyz/">ParallelDots Inc</a>, that can comprehend a huge amount of unstructured text and visual content to empower your products. You can check out some of our <a href="https://www.paralleldots.com/text-analysis-apis">text analysis APIs</a> and reach out to us by filling this form <a href="https://www.paralleldots.com/contact-us">here</a> or write to us at apis@paralleldots.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d57907c44ae6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Niche Applications of Artificial Intelligence in healthcare]]></title>
            <link>https://medium.com/@ParallelDots/niche-applications-of-artificial-intelligence-in-healthcare-9fc7f235e2e?source=rss-62c5679fb60d------2</link>
            <guid isPermaLink="false">https://medium.com/p/9fc7f235e2e</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[ParallelDots]]></dc:creator>
            <pubDate>Fri, 10 Nov 2017 10:57:06 GMT</pubDate>
            <atom:updated>2017-11-10T10:57:06.697Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>“Everything we love about civilization is a product of intelligence, so amplifying our human intelligence with artificial intelligence has the potential of helping civilization flourish like never before — as long as we manage to keep the technology beneficial.“<br>-Max Tegmark, President of the Future of Life Institute</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*aU-eC5u_85hqM0k6.jpg" /></figure><p><strong>Picture credits: iReviews</strong></p><p>Artificial Intelligence has made its way to every field possible, steamrolling the processes along its way. One such field is healthcare. They say healthcare is a field that is not very rules based and a successful doctor is the one who leverages his/her experience to deal with complex and unseen cases. However, there are many low hanging fruits that are already being plucked by AI. This trend is being fueled by increasing digitization in healthcare data and advances in new algorithms. In this piece, we intend to give you a sneak peek into how AI is leading to improved healthcare for humanity. Below are some key examples of research areas and applications.</p><h3>Aiding Histopathology</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PH41a5f3fgRpw9zv.jpg" /></figure><ul><li>Virtual Slides Diagnosis</li><li>Tissue-based diagnosis has seen technological advancement with the introduction of <a href="https://en.wikipedia.org/wiki/Virtual_slide">virtual slides</a>. However, virtual slides demand a lot of time and efforts than that for viewing the original glass slides from the pathologists. This is the time taken in the selection of information containing fields of view. Artificial intelligence can automate the tissue diagnosis routine work. Deep Convolutional Neural Networks are already being used in this area. Automated diagnosis would save a lot of time wasted in supervising and the pathologists can focus on the serious cases.</li><li>Diabetic Retinopathy Treatment</li><li>Diabetic Retinopathy (DR) is the fastest growing cause of blindness, with nearly 415 million diabetic patients at risk worldwide. If not caught early, it can lead to irreversible blindness. In “<a href="http://jamanetwork.com/journals/jama/article-abstract/2588763">Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs</a>“, published by JAMA, Google presented a deep learning algorithm capable of interpreting signs of DR in retinal photographs, potentially helping doctors screen more patients in settings with limited resources.</li><li>Skin Cancer Treatment</li><li>Sebastian Thrun’s lab at Stanford released an <a href="https://www.nature.com/nature/journal/v542/n7639/full/nature21056.html">AI algorithm</a> which detects Skin Cancer with very high accuracy. This algorithm was tested against 21 board-certified dermatologists. In its diagnoses of skin lesions, which represented the most common and deadliest skin cancers, the algorithm matched the performance of professional dermatologists.</li></ul><h3>Medical Diagnosis</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*az8p1QcbpJfdlGVI.jpg" /></figure><p>AI algorithms can aid doctors in medical diagnosis.They can highlight key instances in a person’s previous health history. Incorporating AI into the medical field has the potential to change and vastly improve healthcare in its core. From improved diagnostic accuracy to better-optimized treatment plans, AI could be the key to better medical care for doctors and patients alike.<br>In August 2016, doctors at a hospital in Japan misidentified a 60-year-old woman’s leukemia. But IBM’s Watson examined a vast database of 20 million research papers and made a successful diagnosis in just 10 minutes. The AI-based system can be utilized to prune out the irrelevant data and help the doctor think more clearly focusing on the vital data.</p><h3>Risk Prediction</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*_SmEldLQhMOr2kX3.jpg" /></figure><p>The team of primary care researchers and computer scientists compared a set of standard guidelines from the American College of Cardiology (ACC) with four ‘machine-learning’ algorithms. These algorithms analyzed large amounts of data and self-learn patterns within the data to make predictions on future events which were a patient’s future risk of having heart disease or a stroke, in this case.<br>The results, published in the online journal <a href="http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0174944#sec005">PLOS ONE</a>, showed that the self-teaching ‘artificially intelligent’ tools were remarkably more accurate in predicting cardiovascular disease than the established guidelines. This technology is a godsend for insurance companies by helping them do a more effective appraisal of health risks of a customer.</p><h3>Radiology</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*hTomsxwUnQMnhjNE.jpg" /></figure><p>Applying AI for Radiology is harder as compared to Histopathology and hence we are yet to see groundbreaking results here. There is, however, a lot of work going on in situations where X-rays, CTs, and MRIs can be analyzed automatically, thereby giving radiologists a quick second opinion to consult with.<br>AI has already been used for Chest X-rays for direct diagnosis. Some of the other areas where AI aids diagnosis significantly is segmenting hip bones and lumbar vertebra for QCT/MRI in osteoporosis screening.<br>A Recent release of <a href="http://langlotzlab.stanford.edu/projects/medical-image-net/">Stanford Medical-ImageNet</a> is likely to start a revolution like what ImageNet did for normal images.</p><h3>Automating Drug Discovery</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NfNV7wdCC69coPp-.jpg" /></figure><p>Discovery of a new drug takes years of research, its launch takes even more time and money. Automating drug discovery through AI can tremendously reduce the cost and time as well.The average biomedical researcher deals with a huge amount of new information every day. It is estimated that the bioscience industry is getting 10,000 new publications uploaded on a daily basis from across the globe and among a huge variety of biomedical databases and journals. So, it becomes impossible for the researcher to process the entire information alone. Artificial Intelligence has a vital role to play in elevating the work of drug development researchers.</p><ul><li>A study published in <a href="http://www.cell.com/cell-chemical-biology/fulltext/S2451-9456(16)30291-4">Cell Chemical Biology</a> reveals a big data-based approach to detecting toxic side effects of a drug before it goes to the expensive clinical testing. In the approach called PrOCTOR, researchers analyze each drug using 48 different features to ascertain its safety for clinical use. The entire process is automated using machine learning.</li><li>A company named <a href="http://benevolent.ai/about-us/">BenevolentBio</a> has been doing research into Amyotrophic Lateral Sclerosis (ALS). The AI they’ve developed incorporated in the company’s Judgement Correlation System (JACS) reviews billions of sentences and paragraphs from scientific research papers and abstracts. JACS then links direct relationships between the data and regulates the data into ‘known facts’. These known facts are used to generate a large number of possible hypotheses using criteria set by the scientists. Based on these hypotheses, possible drugs are discovered. They have already managed to identify two potential drug targets for Alzheimer.</li></ul><h3>Some challenges before we conquer</h3><p>However, it comes with its own set of challenges. Data is distributed across many isolated systems, sometimes not even in digital format. This unstructured, uncollated data is a roadblock to developing world class AI solution. Furthermore, healthcare is a regulated space globally. While regulations are necessary, it makes the access to data harder for innovation. The importance of handling patient data carefully can never be undermined. Seamless integration with the traditional systems also presents a challenge for startups and enterprises across the world.</p><h3>The promising future!</h3><p>Given the vast amount of data, healthcare looks to be the most promising sector that will be disrupted by AI. Given the massive impact that can be created by deploying AI solutions in Healthcare, it’s not surprising that some of the world’s leading tech companies are after it. The impending disruption will completely change the way patient care is delivered globally. We here are very excited about the potential of artificial intelligence in a space that is very close to all of our hearts.<br><strong>Here’s to the better health of the planet!</strong></p><p><a href="https://www.paralleldots.com/"><em>ParallelDots</em></a><em> is an ArtificiaI Intelligence research and Deep Learning startup that provides AI solutions to clients in multiple domains. You can check out some of our </em><a href="https://www.paralleldots.com/text-analysis-apis"><em>text analysis APIs</em></a><em> and reach out to us by filling this form </em><a href="https://www.paralleldots.com/contact-us"><em>here.</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9fc7f235e2e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>