# Kaggle Published articles for Kaggle. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Updates on HEIR, the homomorphic encryption compiler project DevFeed: [Updates on HEIR, the homomorphic encryption compiler project]() Original publisher: [Read original article]() Published: 2026-09-04T18:53:40Z Content type: article Language: en Sources: [Jeremy Kun]() Topics: [homomorphic encryption](), [Compiler](), [Encryption](), [Cryptography](), [Machine learning](), [Inference](), [GitHub](), [bazel](), [Kaggle]() Tags: [bazel](), [ckks](), [compiler](), [cryptography](), [encryption](), [github](), [homomorphic-encryption](), [inference](), [kaggle](), [mathematics](), [ml](), [programming]() ### AI overview This companion article explains HEIR, a homomorphic encryption compiler that converts programs to operate directly on encrypted data. It discusses compiling pre-trained machine-learning models for private inference, describes the repository and setup, and reports an example involving encrypted credit-card fraud detection. ### Source excerpt On 2026-08-14 I published an article on the Google Security blog with an update on HEIR, our homomorphic encryption (HE) compiler. This is a companion article, in which I have no limits on word count or jargon, and I can feel free to be honest. So strap in. Assuming you won't read the linked corporate blog post, HEIR is a compiler that converts an input program to a program that operates directly on encrypted data. ## Lessons From the Leaderboard: What 5,000+ Kagglers Taught Us About Improving AI Reasoning DevFeed: [Lessons From the Leaderboard: What 5,000+ Kagglers Taught Us About Improving AI Reasoning]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-07-14T18:20:32Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Artificial Intelligence](), [Benchmark](), [Chain-of-thought](), [datasets](), [Cloud](), [Google](), [NCCL]() Tags: [agentic-ai-generative-ai](), [ai](), [ai-agent](), [ai-inference](), [benchmark](), [blackwell](), [chain-of-thought](), [cost](), [deep-learning](), [engineering](), [evaluation](), [featured](), [google](), [kaggle](), [lora](), [nemotron](), [nvidia](), [performance](), [pre-trained-foundation-models](), [synthetic](), [techniques]() ### AI overview The article distills lessons from NVIDIA's Nemotron Model Reasoning Challenge, where more than 5,000 Kaggle participants tested ways to improve AI reasoning under shared model, infrastructure, and evaluation constraints. It highlights synthetic chain-of-thought data, trace quality, targeted solvers, validation beyond public leaderboards, and careful training and context-budget management. ### Source excerpt The NVIDIA Nemotron Model Reasoning Challenge invited the Kaggle community to explore a focused question: What techniques can improve reasoning accuracy when... ## Fine-tuning a DistilBERT classifier with numerical and text inputs DevFeed: [Fine-tuning a DistilBERT classifier with numerical and text inputs]() Original publisher: [Read original article]() Author: Paloma Jol Published: 2026-04-10T10:04:33Z Content type: tutorial Language: en Sources: [FreeAgent]() Topics: [Fine-tuning](), [Transformer](), [foundation-models](), [Training AI Models](), [datasets](), [Kaggle](), [scikit-learn]() Tags: [ai](), [aws](), [data](), [data-ml](), [data-science](), [dataset](), [encoding](), [fine-tuning](), [hugging-face](), [kaggle](), [llms](), [machine-learning](), [model](), [models](), [nlp](), [performance](), [text-classification](), [training-data](), [validation]() ### AI overview This tutorial explains how to fine-tune a DistilBERT text classifier while incorporating a numerical feature into the same network. It uses the Kaggle wine reviews dataset, combining review descriptions with price and preprocessing the data for training, validation, and testing. ### Source excerpt Text classification is often done through fine-tuning of a pretrained foundation model with domain-specific data. In FreeAgent we use transformer based models to automatically classify incoming bank transactions. Specifically we use a DistilBERT model that is fine-tuned on hundreds of millions of bank transactions with customer-labelled accounting categories. The model inputs are currently text-based, built from a combination of bank transaction descriptions and amounts. In this post we describe an approach to fine-tuning the DistilBERT model and training the classifier including the numerical amount feature as a single network. ## Measuring progress toward AGI: A cognitive framework DevFeed: [Measuring progress toward AGI: A cognitive framework]() Original publisher: [Read original article]() Author: Ryan Burnell Published: 2026-03-17T16:03:47Z Content type: article Language: en Sources: [Google DeepMind News]() Topics: [ai observability]() Tags: [ai](), [benchmarks](), [evaluation](), [framework](), [hackathon](), [kaggle](), [learning](), [memory](), [none](), [reasoning]() ### AI overview The article introduces a cognitive framework for measuring progress toward AGI. It proposes evaluating AI systems across cognitive tasks, comparing results with human baselines, and maps performance against human distributions; it also announces a Kaggle hackathon to help build evaluations. ### Source excerpt We're introducing a framework to measure progress toward AGI, and launching a Kaggle hackathon to build the relevant evaluations. ## Next generation medical image interpretation with MedGemma 1.5 and medical speech to text with MedASR DevFeed: [Next generation medical image interpretation with MedGemma 1.5 and medical speech to text with MedASR]() Original publisher: [Read original article]() Published: 2026-01-13T20:57:16Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [Medical imaging](), [AI Models](), [asr](), [Generative AI](), [Google](), [hugging face](), [Cloud](), [Kaggle]() Tags: [ai-models](), [asr](), [generative-ai](), [google](), [google-cloud](), [health](), [health-bioscience](), [healthcare](), [hugging-face](), [kaggle](), [machine-intelligence](), [medical-imaging]() ### AI overview Google Research describes MedGemma 1.5 4B, an updated open medical generative AI model with improved support for medical imaging, text, medical records, and 2D images. The article also presents MedASR, an open medical speech-to-text model for dictation that can pair with MedGemma for advanced reasoning. The models are available for research and commercial use through Hugging Face and Vertex AI, with a related medical AI hackathon on Kaggle. ### Source excerpt Generative AI ## Google Research 2025: Bolder breakthroughs, bigger impact DevFeed: [Google Research 2025: Bolder breakthroughs, bigger impact]() Original publisher: [Read original article]() Published: 2025-12-18T23:29:00Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [Google](), [Artificial Intelligence](), [Large Language Model](), [Quantum Computing](), [Algorithm](), [data centers](), [Cloud](), [Bioinformatics](), [Benchmark](), [Kaggle]() Tags: [2025](), [agentic](), [ai](), [benchmarks](), [cloud-data](), [data-centers](), [efficiency](), [energy](), [generative](), [google](), [health](), [infrastructure](), [kaggle](), [llm](), [quantum](), [quantum-computing](), [research](), [resilience](), [science](), [year-in-review]() ### AI overview Google Research's 2025 review describes accelerated progress across AI, generative UI, algorithmic research, agentic models, quantum computing, Earth sciences, genomics, biology, neuroscience, climate resilience, health, and education. It also highlights work improving efficiency, reliability, factuality, and cost and energy use in Google products and cloud infrastructure. ### Source excerpt Year in Review ## FACTS Benchmark Suite: Systematically evaluating the factuality of large language models DevFeed: [FACTS Benchmark Suite: Systematically evaluating the factuality of large language models]() Original publisher: [Read original article]() Author: FACTS team Published: 2025-12-09T11:29:03Z Content type: article Language: en Sources: [Google DeepMind News]() Topics: [LLM evaluation / benchmarking](), [Benchmark](), [Large Language Model](), [Kaggle](), [multimodal]() Tags: [benchmark](), [benchmarks](), [evaluation](), [kaggle](), [language-models](), [large-language-models](), [leaderboard](), [multimodal](), [responsibility-safety](), [search](), [web]() ### AI overview Google DeepMind and Kaggle introduce the FACTS Benchmark Suite, which evaluates the factuality of large language models across internal-knowledge, search, multimodal, and grounded-answering tasks. The suite contains 3,513 publicly available examples, with held-out private sets used to calculate an overall FACTS Score and publish results on a leaderboard. ### Source excerpt Systematically evaluating the factuality of large language models with the FACTS Benchmark Suite. ## Improving Hugging Face Model Access for Kaggle Users DevFeed: [Improving Hugging Face Model Access for Kaggle Users]() Original publisher: [Read original article]() Author: Vincent Roseberry; Meg Risdal; Julien Chaumond; Pedro Cuenca; Vaibhav Srivastav Published: 2025-05-14T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [hugging face](), [Kaggle](), [qwen](), [spaces]() Tags: [add-ons](), [announcement](), [code](), [community](), [evaluation](), [hub](), [hugging-face](), [kaggle](), [model](), [qwen](), [qwen3](), [secrets](), [spaces](), [token]() ### AI overview Kaggle is integrating Hugging Face models into its platform, improving model discovery, navigation, and reuse in Kaggle Notebooks. Public notebooks using Hugging Face models can contribute code examples to Kaggle model pages, while private and gated models continue to require Hugging Face authentication. Support for offline Kaggle competition submissions is still in development. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## The TensorFlow Lite Plugin for Flutter is Officially Available DevFeed: [The TensorFlow Lite Plugin for Flutter is Officially Available]() Original publisher: [Read original article]() Author: TensorFlow Blog (noreply@blogger.com) Published: 2023-08-18T16:00:00Z Content type: news Language: en Sources: [The TensorFlow Blog]() Topics: [Flutter](), [TensorFlow Lite](), [Tensorflow](), [cross-platform](), [Mobile](), [object-detection](), [GitHub](), [Google](), [Kaggle]() Tags: [announcement](), [cross-platform](), [explore](), [flutter](), [github](), [google](), [inference](), [kaggle](), [mobile](), [object-detection](), [plugin](), [tensorflow](), [tensorflow-lite](), [tensorflowlite](), [tutorial]() ### AI overview The TensorFlow Lite plugin for Flutter has been officially migrated to the TensorFlow GitHub account and released. The updated plugin adds features and example apps, including live-camera object detection, and enables local TensorFlow model inference in Flutter apps across supported platforms. ### Source excerpt Posted by Paul Ruiz, Developer Relations Engineer We're excited to announce that the TensorFlow Lite plugin for Flutter has been officially migrated to the TensorFlow GitHub account and released! Three years ago, Amish Garg, one of our talented Google Summer of Code contributors, wrote a widely used TensorFlow Lite plugin for Flutter. The plugin was so popular that we decided to migrate it to our official repo, making it easier to maintain directly by the Google team. We are grateful to Amish for his contributions to the TensorFlow Lite Flutter plugin. Through the efforts of developers in the community, the plugin has been updated to the latest version of TensorFlow Lite, and a collection of new features and example apps have been added, such as object detection through a live camera feed. So what is TensorFlow Lite? TensorFlow Lite is a way to run TensorFlow models on devices locally, supporting mobile, embedded, web, and edge devices. TensorFlow Lite's cross-platform support and on-device performance optimizations make it a great addition to the Flutter development toolbox. Our goal with this plugin is to make it easy to integrate TensorFlow Lite models into Flutter apps across mobile platforms, with desktop support currently in development through the efforts of our developer community. Find pre-trained TensorFlow Lite models on model repos like Kaggle Models or create your own custom TensorFlow Lite models. Let's take a look at how you could use the Flutter TensorFlow Lite plugin for image classification: TensorFlow Lite Image Classification with Flutter First you will need to install the plugin from pub.dev. Once the plugin is installed, you can load a TensorFlow Lite model into your Flutter app and define the input and output tensor shapes. If you're using the MobileNet model, then the input tensor will be a 224 by 224 RGB image, and the output will be a list of confidence scores for the trained labels. // Load model Future _loadModel() async { final opt ## Hand Gesture Recognition on ESP32-S3 with ESP-Deep Learning DevFeed: [Hand Gesture Recognition on ESP32-S3 with ESP-Deep Learning]() Original publisher: [Read original article]() Author: John Lee Published: 2022-12-02T00:00:00Z Content type: tutorial Language: en Sources: [Blog on Developer Portal]() Topics: [ESP32-S3](), [Deep learning](), [Artificial Intelligence](), [ESP-IDF](), [datasets](), [Model Development](), [quantization]() Tags: [ai](), [blog](), [classification](), [deep-learning](), [esp-dl](), [esp-id](), [esp32](), [esp32-s3](), [espressif](), [kaggle](), [quantization]() ### AI overview This tutorial explains how to deploy a deep-learning model for classifying six hand gestures on the ESP32-S3 using Espressif's ESP-DL framework. It covers prerequisites, model development, dataset preparation, train/test/calibration splits, and model quantization. ### Source excerpt Artificial intelligence transforms the way computers interact with the real world. Decisions are carried by getting data from Tiny low-powered devices and sensors into the cloud. Connectivity, high cost and data privacy are some of the demerits of this method. Edge artificial intelligence is another way to process the data right on the physical device without sending data back and forth improving the latency and security and reducing the bandwidth and power. ## UK Bank Holidays with Pandas on Jupyter DevFeed: [UK Bank Holidays with Pandas on Jupyter]() Original publisher: [Read original article]() Author: Simone Spaccarotella Published: 2022-09-21T17:22:24Z Content type: tutorial Language: en Sources: [BBC]() Topics: [pandas](), [Jupyter Notebook](), [JSON](), [Python](), [Data Science](), [data]() Tags: [coding](), [data-science](), [json](), [jupyter-notebook](), [kaggle](), [pandas](), [python]() ### AI overview A hands-on tutorial about using Pandas in a Jupyter Notebook to retrieve UK bank holiday data in JSON format and prepare it for analysis in a DataFrame. The article also describes the author's self-study practice in Python, data science, and machine learning. ### Source excerpt How I learned to manipulate JSON data with Pandas on a Jupyter Notebook and deconstruct it to a DataFrame ready for queries.Image by author created from Jupiter photo by NASA and Pandas photo by Pascal Müller on UnsplashA bit of context first I started a self-study path to learn the theoretical fundamentals of Data Science and Machine Learning. I have also been playing with Python ever since, with coding exercises during the "Friday 10% time" here at the BBC and on my own time when I have some to spare. I created this playground repository to refresh my Pythonic knowledge that was buried deep inside the funny head of mine. This repo is a mix of material and exercises to keep my Python coding skills sharp. GitHub - simonespa/python-playground: A playground repo for Python I also created a specific playground focused on Data Science and Machine learning. In this playground I practice with libraries such as NumPy, Pandas, Scikit-Learn, TensorFlow (et. all) with a Jupyter Notebook. It also contains notes and summaries of what I learn in forms of README or notebooks. GitHub - simonespa/machine-learning-playground: A playground to practice machine learning, data modelling and visualisation Nothing fancy or ground-breaking, just a couple of playground repos for me to practice and learn at my own pace. The 10% time at the BBC This happens every other Friday and it is about learning. Engineers use this time to learn new technologies and/or languages, discover new approaches and techniques. We can organise the day as we want. It could be a self-paced learning session (reading, watching tutorials, etc.), a day of study for an apprenticeship or a university degree etc., an actual hands-on prototyping session to work on a long-running personal project, or a group collaboration with other like-minded people. There is no limitation to how we plan our day, as long as we keep in mind the end goal. In simple terms, satisfy our own curiosity and creativity by strengthening our technic ## Announcing Our Market Prediction Kaggle Competition DevFeed: [Announcing Our Market Prediction Kaggle Competition]() Original publisher: [Read original article]() Author: Craig Falls Published: 2020-11-24T00:00:00Z Content type: release Language: en Sources: [Jane Street]() Topics: [Kaggle](), [Machine learning](), [data](), [Data Science]() Tags: [competition](), [data](), [data-science](), [financial](), [hiring](), [kaggle](), [london](), [machine-learning](), [ml]() ### AI overview Jane Street announces a Kaggle competition based on a real problem and real financial data, with $100,000 in total prizes. The company says the contest aims to introduce more people to its data-science work and encourage applications. ### Source excerpt Jane Street is running a Kaggle contest based on a real problem with real financial data. If you like ML projects, or think you might, head over and check it out. We think it's a pretty fun one. The prizes are pretty good too, with a total $100K being paid out. ## Machine Learning Datasets for NLP DevFeed: [Machine Learning Datasets for NLP]() Original publisher: [Read original article]() Author: Fuzzygroup Published: 2020-03-09T00:00:00Z Content type: article Language: en Sources: [Scott Johnson]() Topics: [datasets](), [Natural language processing](), [Machine learning](), [data](), [Kaggle]() Tags: [data](), [datasets](), [kaggle](), [machine-learning](), [netlabeler](), [nlp]() ### AI overview A short list of machine-learning datasets for natural language processing, including resources from Kaggle, LionBridge.ai, the Million Song Dataset, Synced Review, Twitter, and Sample Stream. ### Source excerpt Today is a short one, just pointers to some datasets for Machine Learning data sets that I keep losing. Kaggle LionBridge.ai Million Song Dataset Synced Review Twitter Sample Stream ## Preparing your product for machine learning DevFeed: [Preparing your product for machine learning]() Original publisher: [Read original article]() Author: Joakim Rishaug Published: 2018-08-09T08:00:00Z Content type: tutorial Language: en Sources: [Finn.no]() Topics: [Machine Learning & Artificial Intelligence](), [data](), [datasets](), [Google](), [Kaggle]() Tags: [data](), [introduction](), [kaggle](), [machine-learning]() ### AI overview This article explains how engineering teams can prepare products for future machine learning work, even when data scientists are not yet available. It emphasizes engineering fundamentals, useful features, numerical data, and reducing problems caused by unstructured or incomplete data. ### Source excerpt At many companies, there are few data scientists and many projects which may be very interesting for the business to apply machine learning on. For most of these companies, starting greenfield projects with a data scientist on the team might be either a difficult political battle (to get assigned resources) or impossible, if there are no data scientists yet. But how can a team prepare their project for the day when there will be a data scientist around, and in the process, maybe improve their current product as well? In Google's "Rules of ML", they make some good suggestions for this in their introduction: Do machine learning like the great engineer you are, not the great machine learning expert you aren't. Most of the problems you will face, are in fact engineering problems. Most of the gains come from great features, not great machine learning algorithms. This means, the team already has great potential to improve their product in ways that the data scientist might help them do in a more structured way later. I believe as long as the engineers and programmers know how machine learning algorithms like having the data shaped, they can take advantage of this to move forward confidently. What kind of data does a machine learning algorithm like? Computers like things that are quantifiable by numbers, and this is probably obvious to many programmers, but it's easy to forget this when we have tools like Google that can find and seemingly understand what we want just by typing a couple of words. According to a Kaggle survey from 2017, "dirty data" is the biggest problem faced by machine learning practitioners. This means data being unstructured in some form, either by being severely skewed in one direction or the other, being full of holes (only subsets of users even have a given feature), or data that has to go through processing before actually becoming a usable feature. Making sure that this friction is minimized is important later for the progress of machine learning