Machine Learning
9 pieces on Machine Learning, including 2 step-by-step guides.
Guides
Top 5 Small Language Models for Agentic Tool Calling
Explore 5 small language models (SLMs) that feature open weights and first-class tool-calling support for local agentic workflows.
Learning to Rank: The Hidden Layer Powering Modern RAG
Your RAG pipeline is only as good as your ranker. Explore the technical depths of Learning to Rank (LTR), from LambdaMART's 'Physics Trick' to the Cross-Encoder tradeoff that determines production success.
News & Analysis
WeatherNext 3 Hits 5 km Resolution With Hourly Satellite Initialization
Google DeepMind's WeatherNext 3 resolves forecasts to 0.05° (~5 km), re-initializes every hour from live satellite data, and cuts precipitation CRPS up to 60%.
WeatherNext 3 Delivers 5x Sharper Forecasts via Live Satellite Data
Google DeepMind's WeatherNext 3 replaces NWP training data with live geostationary satellite mosaics, cutting precipitation error by up to 60% and refreshing hourly at 5 km resolution.
Swiggy's 135K-Parameter MLP Beats Third-Party pLTV With 350+ Features
Swiggy's four-head MLP cut parameters 63% to 135,000 while improving accuracy, outperforming a third-party pLTV platform on retention and gross order value.
Gemini 3.5 Transcribe: 2.6% WER, 85+ Languages, Two API Surfaces
Google's Gemini 3.5 Transcribe posts 2.6% non-streaming WER across 85+ languages, with a dual-endpoint design that forces hard architectural choices.
Google HEIR Compiles PyTorch Models for Fully Homomorphic Encryption
Google's open-source HEIR toolchain compiles pre-trained PyTorch models for FHE inference, with ~10³ overhead and no published LLM benchmarks yet.
SigLIP LoRA Fine-Tune Cuts Under-Labeling from 9.4% to 3.4%
Alma Media's 23-class image classifier shows why under-labeling rate, not F1, is the right signal for deciding whether LoRA fine-tuning is worth the cost.
WhatsApp Scam Alert Runs ML On-Device with Confidential VM Telemetry
WhatsApp's beta Scam Alert feature classifies messages locally, routes telemetry through confidential VMs with differential privacy, and verifies models via a third-party transparency ledger.