https://samarthya.me/categories/ml/Recent content in Ml on SamarthyaHugo -- gohugo.ioen© Samarthya 2018-2026. All rights reserved.Sat, 31 Jan 2026 12:00:00 +0000https://samarthya.me/posts/language-model/Sat, 31 Jan 2026 12:00:00 +0000https://samarthya.me/posts/language-model/<figure><img class="my-0 rounded-md" loading="lazy" decoding="async" fetchpriority="auto" alt="" width="600" height="600" src="https://samarthya.me/posts/language-model/transformer1.png" srcset="https://samarthya.me/posts/language-model/transformer1.png 800w, https://samarthya.me/posts/language-model/transformer1.png 1280w" sizes="(min-width: 768px) 50vw, 65vw" data-zoom-src="https://samarthya.me/posts/language-model/transformer1.png"></figure> <p>Think of a language model as a really sophisticated autocomplete system. You know how your phone suggests the next word when you’re texting? A language model works on that same principle, but at a much more advanced level.</p>https://samarthya.me/posts/matrix-multiplication-to-gpus/Thu, 16 Oct 2025 12:00:00 +0000https://samarthya.me/posts/matrix-multiplication-to-gpus/<figure><img class="my-0 rounded-md" loading="lazy" decoding="async" fetchpriority="auto" alt="" width="600" height="600" src="https://samarthya.me/posts/matrix-multiplication-to-gpus/ai-1.png" srcset="https://samarthya.me/posts/matrix-multiplication-to-gpus/ai-1.png 800w, https://samarthya.me/posts/matrix-multiplication-to-gpus/ai-1.png 1280w" sizes="(min-width: 768px) 50vw, 65vw" data-zoom-src="https://samarthya.me/posts/matrix-multiplication-to-gpus/ai-1.png"></figure> <p>If you are like me, you must have stumbled on this question – Why is matrix multiplication so significant – especially in the AI/ML universe.</p>https://samarthya.me/posts/the-core-divide/Fri, 10 Oct 2025 12:00:00 +0000https://samarthya.me/posts/the-core-divide/<figure><img class="my-0 rounded-md" loading="lazy" decoding="async" fetchpriority="auto" alt="" width="600" height="600" src="https://samarthya.me/posts/the-core-divide/ml-1.png" srcset="https://samarthya.me/posts/the-core-divide/ml-1.png 800w, https://samarthya.me/posts/the-core-divide/ml-1.png 1280w" sizes="(min-width: 768px) 50vw, 65vw" data-zoom-src="https://samarthya.me/posts/the-core-divide/ml-1.png"></figure> <p>At the heart of machine learning, algorithms learn from data. The key distinction between supervised and unsupervised learning is the <strong>type of data</strong> used for training.</p>https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/Thu, 09 Oct 2025 12:00:00 +0000https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/<figure><img class="my-0 rounded-md" loading="lazy" decoding="async" fetchpriority="auto" alt="" width="600" height="600" src="https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/arrow-2.png" srcset="https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/arrow-2.png 800w, https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/arrow-2.png 1280w" sizes="(min-width: 768px) 50vw, 65vw" data-zoom-src="https://samarthya.me/posts/beyond-the-dataframe-how-parquet-and-arrow-turbocharge-pyspark/arrow-2.png"></figure> <p>In my <a href="https://samarthya.me/posts/pyspark-vs-pandas-choosing-right/" >last post</a>, we explored the divide between <strong>Pandas</strong> (single machine) and <strong>PySpark</strong> (distributed computing).</p> <blockquote><p>The conclusion: for massive datasets, PySpark is the clear winner.</p> </blockquote>