<?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 Georgia Deaconu on Medium]]></title>
        <description><![CDATA[Stories by Georgia Deaconu on Medium]]></description>
        <link>https://medium.com/@georgiadeaconu?source=rss-c4a98f38b0e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*-JLMC1GeauTqpwbaxWmzzA.jpeg</url>
            <title>Stories by Georgia Deaconu on Medium</title>
            <link>https://medium.com/@georgiadeaconu?source=rss-c4a98f38b0e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 02:29:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@georgiadeaconu/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[Inside the Machine’s Mood: Exploring the Emotional Circuits that drive AI Behavior]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://pub.towardsai.net/inside-the-machines-mood-exploring-the-emotional-circuits-that-drive-ai-behavior-5d302177b098?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/948/1*hjiJT2r3Fe_f1MI9r-oU0g.png" width="948"></a></p><p class="medium-feed-snippet">Anyone who has interacted with an LLM knows there is an emotional component to the interaction. LLMs are &#x201C;happy&#x201D; to assist you with your&#x2026;</p><p class="medium-feed-link"><a href="https://pub.towardsai.net/inside-the-machines-mood-exploring-the-emotional-circuits-that-drive-ai-behavior-5d302177b098?source=rss-c4a98f38b0e------2">Continue reading on Towards AI »</a></p></div>]]></description>
            <link>https://pub.towardsai.net/inside-the-machines-mood-exploring-the-emotional-circuits-that-drive-ai-behavior-5d302177b098?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5d302177b098</guid>
            <category><![CDATA[ai-safety]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[emotions]]></category>
            <category><![CDATA[ai-alignment]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Thu, 16 Apr 2026 23:31:01 GMT</pubDate>
            <atom:updated>2026-04-22T08:44:40.500Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Did AI coding assistants change Software Engineering?]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://levelup.gitconnected.com/did-ai-coding-assistants-change-software-engineering-5efb21cb3d29?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/1024/1*yzjp9CcjCsaqml5BmkJYiQ.png" width="1024"></a></p><p class="medium-feed-snippet">It&#x2019;s been more than 2 years since ChatGPT became the first AI system to gain widespread adoption. People quickly realized that the&#x2026;</p><p class="medium-feed-link"><a href="https://levelup.gitconnected.com/did-ai-coding-assistants-change-software-engineering-5efb21cb3d29?source=rss-c4a98f38b0e------2">Continue reading on Level Up Coding »</a></p></div>]]></description>
            <link>https://levelup.gitconnected.com/did-ai-coding-assistants-change-software-engineering-5efb21cb3d29?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5efb21cb3d29</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Fri, 27 Dec 2024 02:31:20 GMT</pubDate>
            <atom:updated>2024-12-27T16:08:55.954Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Let NotebookLM Tell Your Story (While You Sit Back and Enjoy the Praise)]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://aiadvances.org/let-notebooklm-tell-your-story-while-you-sit-back-and-enjoy-the-praise-a1a6ce4c8ad4?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/1425/1*9rh2ISNw0Wpba8503sYGRg.png" width="1425"></a></p><p class="medium-feed-snippet">Or how to use AI to cheer you up in moments of doubt</p><p class="medium-feed-link"><a href="https://aiadvances.org/let-notebooklm-tell-your-story-while-you-sit-back-and-enjoy-the-praise-a1a6ce4c8ad4?source=rss-c4a98f38b0e------2">Continue reading on AI Advances »</a></p></div>]]></description>
            <link>https://aiadvances.org/let-notebooklm-tell-your-story-while-you-sit-back-and-enjoy-the-praise-a1a6ce4c8ad4?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/a1a6ce4c8ad4</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[notebooklm]]></category>
            <category><![CDATA[fun]]></category>
            <category><![CDATA[gen-ai-tools]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Sun, 06 Oct 2024 13:03:00 GMT</pubDate>
            <atom:updated>2024-10-18T16:11:22.551Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[How to use InstantID to generate identity preserving images]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://aiadvances.org/how-to-use-instantid-to-generate-identity-preserving-images-f6fe126277fc?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/2600/1*stkNStfCdiOQZx75KcgQAw.png" width="3656"></a></p><p class="medium-feed-snippet">Or how to generate better pictures of yourself</p><p class="medium-feed-link"><a href="https://aiadvances.org/how-to-use-instantid-to-generate-identity-preserving-images-f6fe126277fc?source=rss-c4a98f38b0e------2">Continue reading on AI Advances »</a></p></div>]]></description>
            <link>https://aiadvances.org/how-to-use-instantid-to-generate-identity-preserving-images-f6fe126277fc?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/f6fe126277fc</guid>
            <category><![CDATA[gen-ai-tools]]></category>
            <category><![CDATA[stable-diffusion]]></category>
            <category><![CDATA[image-generation]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Tue, 13 Feb 2024 13:26:46 GMT</pubDate>
            <atom:updated>2024-02-15T08:31:54.538Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[LLM Safety Training and Jail-Breaking]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://aiadvances.org/llm-safety-training-and-jail-breaking-57417f486d9f?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/600/1*4CNKlYcO3maRMTbqEUBNJw.png" width="600"></a></p><p class="medium-feed-snippet">Since ChatGPT brought LLMs into the spotlight, we witnessed the cat and mouse game between LLM developers trying to implement guardrails&#x2026;</p><p class="medium-feed-link"><a href="https://aiadvances.org/llm-safety-training-and-jail-breaking-57417f486d9f?source=rss-c4a98f38b0e------2">Continue reading on AI Advances »</a></p></div>]]></description>
            <link>https://aiadvances.org/llm-safety-training-and-jail-breaking-57417f486d9f?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/57417f486d9f</guid>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[genai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-safety]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Sat, 20 Jan 2024 00:32:17 GMT</pubDate>
            <atom:updated>2026-04-22T08:47:28.866Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Towards LLM explainability: why did my model produce this output ?]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/data-science/towards-llm-explainability-why-did-my-model-produce-this-output-8f730fc73713?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/1595/1*bMAaz_buS5DG7KPfXBPDzg.png" width="1595"></a></p><p class="medium-feed-snippet">The recent release of larger, better LLMs that showcase new capabilities has been paired with growing concerns over AI safety</p><p class="medium-feed-link"><a href="https://medium.com/data-science/towards-llm-explainability-why-did-my-model-produce-this-output-8f730fc73713?source=rss-c4a98f38b0e------2">Continue reading on TDS Archive »</a></p></div>]]></description>
            <link>https://medium.com/data-science/towards-llm-explainability-why-did-my-model-produce-this-output-8f730fc73713?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/8f730fc73713</guid>
            <category><![CDATA[explainable-ai]]></category>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[ai-safe]]></category>
            <category><![CDATA[ai-safety]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Fri, 15 Dec 2023 00:50:50 GMT</pubDate>
            <atom:updated>2026-04-22T08:39:51.293Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Display Spark logs from AWS Glue jobs in a local Spark History Server]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-snippet">With AWS Glue it is easy to create and run Spark ETL jobs without having to think about the underlying infrastructure. Unfortunately it&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/@georgiadeaconu/display-spark-logs-from-aws-glue-jobs-in-a-local-spark-history-server-4a539e13343a?source=rss-c4a98f38b0e------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@georgiadeaconu/display-spark-logs-from-aws-glue-jobs-in-a-local-spark-history-server-4a539e13343a?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/4a539e13343a</guid>
            <category><![CDATA[aws-glue]]></category>
            <category><![CDATA[data-engineering]]></category>
            <category><![CDATA[logging-and-monitoring]]></category>
            <category><![CDATA[metrics-and-analytics]]></category>
            <category><![CDATA[spark]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Sat, 19 Aug 2023 06:03:15 GMT</pubDate>
            <atom:updated>2023-09-01T12:28:14.262Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[5 things you never wanted to know about the Eurovision]]></title>
            <link>https://medium.com/@georgiadeaconu/5-things-you-never-wanted-to-know-about-the-eurovision-3bdf6579a45c?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/3bdf6579a45c</guid>
            <category><![CDATA[eurovision]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[fun]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Wed, 17 May 2023 21:30:42 GMT</pubDate>
            <atom:updated>2026-04-22T09:11:24.707Z</atom:updated>
            <content:encoded><![CDATA[<h3>5 Things You Never Wanted to Know about the Eurovision Song Contest</h3><p>The <a href="https://en.wikipedia.org/wiki/Eurovision_Song_Contest">Eurovision Song Contest</a> has been held annually since 1956 and it is currently the longest-running annual international televised music competition in the world. This year’s competition took place on the 13th of May and was won by Sweden (again!).</p><p>I have been watching it quite regularly since 2008 and I think I have accumulated some potentially biased opinions about it. The following analysis is based on the <a href="https://www.kaggle.com/datasets/datagraver/eurovision-song-contest-scores-19752019">Kaggle dataset</a> containing the detailed voting results from 1975 to 2019. The <a href="https://en.wikipedia.org/wiki/Voting_at_the_Eurovision_Song_Contest">voting system</a> is a bit complex but somehow to me it always seemed to reflect some of the geopolitical realities involving the competing countries. I wanted to see if any of it is verified by the available data.</p><p>The full code for the analysis is available <a href="https://github.com/gedec/Eurovision_analysis/blob/main/eurovision.ipynb">here</a>.</p><h3>1. Does anyone still care about the Eurovision?</h3><p>Given the available data, I have assessed the contest’s popularity based on the number of countries that are willing to participate (choosing a candidate is usually quite tedious and can spark nationwide debates). The contest’s popularity seems to have increased in the last 15 years, with around 40 countries participating in recent years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/968/1*mISLK_-7sXhiGkORiMl6Yw.png" /><figcaption>Image by the author</figcaption></figure><p>52 countries participated overall (including some countries that no longer exist, such as Yugoslavia), out of which the top 10 most constant participants are:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/470/1*jeTL2Yg6gC2qIPE68dpuSw.png" /><figcaption>Image by the author</figcaption></figure><h3>2. Has Sweden really won that often?</h3><p>This impression of mine is easy to check because I can count the number of times each country has won. Turns out Sweden does not have the most wins, it has won 4 times over the studied period and this year’s win puts it at 5 wins overall. This still means it is the country with the second-most wins overall.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/887/1*WlzhWE4QrydTi7r42d_y6g.png" /><figcaption>Image by the author</figcaption></figure><h3>3. Does geography influence the probability of winning?</h3><p>To create the following geoplot I followed the steps described in this article about<a href="https://towardsdatascience.com/geoplotting-emissions-intensity-of-electricity-generation-in-europe-90c22b378858"> Geoplotting emissions of electricity generation</a>. This plot shows that most wins are concentrated in Northern Europe, but that most countries had at least one win (sorry Australia, it was too hard to include you in this plot without making it look weird).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/1*Zqxf-j1MZriqLAYOg2yhbg.png" /><figcaption>Image by the author</figcaption></figure><h3>4. Is France really that unpopular?</h3><p>I have been watching the Eurovision mostly on French television, and each year, the moment when they count the votes seems very painful for the presenters. Even if France has won 2 times, when it doesn’t win, it seems like the overall result is really bad.</p><p>To understand if France is really that unpopular, I took a look at the sum of awarded points per year for each country and calculated the percentile corresponding to France’s result. The overall distribution of the points varies each year, but the percentile should be a good indicator of France’s overall position.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jZKy7mSRsl6oOsvdudHMOw.png" /><figcaption>Image by the Author</figcaption></figure><p>When I look at the computed values, they tell a different story than my previous impressions. It is true that France seems to have had a few rough years at the Eurovision, and that when it is bad, it is bad for at least 3 years in a row.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/775/1*HzGvQ0ZJzz0mA6SFMD499g.png" /></figure><p>But overall, France’s performance seems to be quite uniformly distributed, with no obvious trend over the years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*5rHECMFBfBjPx6b7SwWwJA.png" /></figure><h3>5. Do some countries consistently vote for their neighbors/friends?</h3><p>To answer this question, I computed for each country the sum of the points awarded by each of the other countries over the studied period. The results tell an interesting story.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/343/1*47KrQhxWDT-Gkk2oJWtlwA.png" /><figcaption>Sum off points awarded from country to country (Image by the Author)</figcaption></figure><p>We can see that Cyprus was very consistent in awarding its points to Greece and that Greece returned the favor (a bit of a political message there). The Nordic countries tend to vote for each other (looking at you Norway, Sweden, Finland, Denmark and Iceland). The UK mostly votes for Ireland, and Moldova consistently votes for Romania (when Romania manages to make it to the final). Romania returns the favor most of the time.</p><p>Some immigration based patterns could explain the fact that France seems to mostly vote for Portugal and that Germany mostly votes for Turkey. Spain votes for the neighboring Portugal but also for Greece (I cannot really explain this one). This seems to confirm some of my intuitions, but not in a very spectacular manner.</p><p>This view also confirms France’s lack of overall popularity amongst fellow Europeans, in the sense that there is no country that consistently votes for France, regardless of how good their song is. In that sense, France’s results generally reflect the quality of the song, as it does not benefit from any obvious sympathy voting (unlike Greece and the Nordic countries).</p><h3>Conclusions</h3><p>It turns out that people still care about the Eurovision song contest and that I was wrong in thinking that Sweden wins very often. There seems to be something special about Northern Europe that makes it come up with better songs, but the countries are also consistently voting for each other.</p><p>France does not seem to be as unpopular as I thought, even if in the last 20 years it seems to have alternated very good years with very bad years. History and immigration patterns do seem to influence a country’s votes, although my analysis didn’t go in too much depth on the topic.</p><p>I hope you had as much fun reading this as I had writing it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3bdf6579a45c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Data Leakage affects model performance claims]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/data-science/how-data-leakage-affects-model-performance-claims-841eb01276bb?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/600/1*WpS9eIQ86_3GqX3yOLce9w.png" width="600"></a></p><p class="medium-feed-snippet">This year has seen several important scientific advancements enabled by machine learning driven research. Along with the enthusiasm came&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/data-science/how-data-leakage-affects-model-performance-claims-841eb01276bb?source=rss-c4a98f38b0e------2">Continue reading on TDS Archive »</a></p></div>]]></description>
            <link>https://medium.com/data-science/how-data-leakage-affects-model-performance-claims-841eb01276bb?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/841eb01276bb</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[data-leakage-prevention]]></category>
            <category><![CDATA[data-quality]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Mon, 02 Jan 2023 18:16:20 GMT</pubDate>
            <atom:updated>2023-01-17T08:30:56.089Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Monitoring Databricks jobs through calls to the REST API]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://medium.com/data-science/monitoring-databricks-jobs-through-calls-to-the-rest-api-4c02d7d27278?source=rss-c4a98f38b0e------2"><img src="https://cdn-images-1.medium.com/max/666/1*xR2yv57K2hNVtk5vIvvadw.png" width="666"></a></p><p class="medium-feed-snippet">Monitoring jobs that run in a Databricks production environment requires not only setting up alerts in case of failure but also being able&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/data-science/monitoring-databricks-jobs-through-calls-to-the-rest-api-4c02d7d27278?source=rss-c4a98f38b0e------2">Continue reading on TDS Archive »</a></p></div>]]></description>
            <link>https://medium.com/data-science/monitoring-databricks-jobs-through-calls-to-the-rest-api-4c02d7d27278?source=rss-c4a98f38b0e------2</link>
            <guid isPermaLink="false">https://medium.com/p/4c02d7d27278</guid>
            <category><![CDATA[monitoring]]></category>
            <category><![CDATA[mlops]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[databricks]]></category>
            <dc:creator><![CDATA[Georgia Deaconu]]></dc:creator>
            <pubDate>Sun, 09 Oct 2022 21:24:38 GMT</pubDate>
            <atom:updated>2022-10-11T21:13:58.715Z</atom:updated>
        </item>
    </channel>
</rss>