<?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 Santiago Becerra on Medium]]></title>
        <description><![CDATA[Stories by Santiago Becerra on Medium]]></description>
        <link>https://medium.com/@santibecerra?source=rss-e7b5d44354a6------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*QCmw6ntnpUPtmSVPxfHDzg.jpeg</url>
            <title>Stories by Santiago Becerra on Medium</title>
            <link>https://medium.com/@santibecerra?source=rss-e7b5d44354a6------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 19:03:44 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@santibecerra/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[Using machine learning to make sense of marine radio communications — Part One : Why?]]></title>
            <link>https://medium.com/@santibecerra/using-machine-learning-to-make-sense-of-marine-radio-communications-part-one-why-6d5fbe00688?source=rss-e7b5d44354a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/6d5fbe00688</guid>
            <dc:creator><![CDATA[Santiago Becerra]]></dc:creator>
            <pubDate>Tue, 23 Jan 2024 16:46:56 GMT</pubDate>
            <atom:updated>2024-01-23T16:46:56.933Z</atom:updated>
            <content:encoded><![CDATA[<h3>Using machine learning to make sense of marine radio communications — Part One : Why?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HGMPWYADWpxhiCeINbWkNg.jpeg" /></figure><p>As a sailor who frequents the busy waters off southern California, I rely on a standard suite of electronics to help me navigate safely. On most boats, these include a GPS/chart-plotter, radar, and VHF radio. With the GPS/chart-plotter, I see my position, heading, and speed on a detailed map which shows land masses, water depths, and other stationary obstructions such as buoys and shallow reefs. The radar overlays above-water obstructions in a 36 mile radius around my position such as raised land masses, other vessels, or areas of dense rain. With the VHF radio, I can transmit and receive audio within a 20–40 mile radius around my current position. With an additional AIS receiver, I can also use the VHF radio to pickup position information transmitted digitally from nearby vessels equipped with an AIS transponder (which most military and commercial ships have). This info can get sent to the chart plotter automatically.</p><p>It is worth noting that despite the proliferation of cellular and satellite phones on vessels, VHF radio remains the primary voice communication (and safety) device for all boaters. Unlike phones which allow one device to call another device, VHF radio designates a set of channels on which any radio can broadcast or listen. Channel 16 is the internationally designated distress frequency, secondarily used to establish communications for routine calls before switching to another channel. All vessels carrying a VHF radio are required by maritime law to have it always tuned to Channel 16. The US Coast Guard maintains sensitive antennas and monitors this channel as well.</p><p>On any given day, while sailing in busy near-shore waters you will hear a variety of radio broadcasts, many of which will require some level of attention. They can involve things such as: a vessel in distress, a vessel alerting other craft in the the area of an unmarked obstruction, or a vessel hailing another vessel on a collision course (they might be talking to you!).</p><p>It is critical for boaters to determine whether each of these broadcasts requires an action on their part. I have found that many recreational boaters are often too busy fishing, socializing, or tending to other activities to properly make this determination.</p><p>Consider the following radio broadcast :</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fclyp.it%2F3kirrjeq%2Fwidget&amp;url=https%3A%2F%2Fclyp.it%2F3kirrjeq&amp;image=https%3A%2F%2Fstatic.clyp.it%2Fsite%2Fimages%2Flogos%2Fclyp-og-1200x630.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=clyp" width="720" height="265" frameborder="0" scrolling="no"><a href="https://medium.com/media/50547b60ff74d72c30d01832a63098de/href">https://medium.com/media/50547b60ff74d72c30d01832a63098de/href</a></iframe><p>If you are a boater out on the water and you hear this broadcast, it should catch your attention! A navy ship located at “32 degrees 43 minutes North 117 degrees 27 minutes West” is conducting “manned autonomy testing” and advising vessels to keep a 3 mile distance. Even with a chart-plotter and GPS, you would still have to scramble to find a way to jot this down, then go your your plotter and either enter the coordinates or eyeball the location to find out where the vessel is and whether you are close enough to it to alter course. Keep in mind, these broadcasts are often only sent once, further increasing the likelihood of a busy captain to miss the chance to jot down key information.</p><p>Persistently faced with this annoying and potentially dangerous problem, I started thinking about how we could use machine learning to help make sense of VHF voice communications, with the goal of providing actionable data that makes boating safer. In the context of the problem described above, could we automatically detect when coordinates are being spoken on the radio, and map that location directly on chart-plotter? Could we make additional sense of the broadcast such as detecting the severity and proximity and surface that information?</p><p>Taking a broader perspective, voice radio communications are employed beyond the marine environment. Police, firefighters and other organizations involved in any form of distress response rely on voice radio communications. A system which could scalably capture and process this information in real-time could have tremendous value in terms of providing real-time insights to first responders. Moreover, it could provide a significant corpus of data, which could be used to analyze patterns or conduct forensics on significant events.</p><p>I decided to build a simple POC that would tackle the narrow problem of detecting and plotting coordinates spoken over VHF channel 16 on a map. While this functionality could surface as a feature of a VHF radio or a mobile app, the POC also keeps an eye on how one could build a system for the broader vision outlined above.</p><p>Stay tuned for the next installment which will describe the POC!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6d5fbe00688" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Simple neural network implementation in C]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/data-science/simple-neural-network-implementation-in-c-663f51447547?source=rss-e7b5d44354a6------2"><img src="https://cdn-images-1.medium.com/max/2600/0*W1ohPWGX7T48uVTh" width="4032"></a></p><p class="medium-feed-snippet">I decided to write a simple neural network from scratch in C, without the help of any vector or matrix libraries. Why would I do this?</p><p class="medium-feed-link"><a href="https://medium.com/data-science/simple-neural-network-implementation-in-c-663f51447547?source=rss-e7b5d44354a6------2">Continue reading on TDS Archive »</a></p></div>]]></description>
            <link>https://medium.com/data-science/simple-neural-network-implementation-in-c-663f51447547?source=rss-e7b5d44354a6------2</link>
            <guid isPermaLink="false">https://medium.com/p/663f51447547</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[neural-networks]]></category>
            <dc:creator><![CDATA[Santiago Becerra]]></dc:creator>
            <pubDate>Thu, 03 Oct 2019 08:04:35 GMT</pubDate>
            <atom:updated>2021-05-14T22:17:37.202Z</atom:updated>
        </item>
    </channel>
</rss>