Machine Learning Research - The Batch

Kids in costumes face a locked door, with candy visible behind bars, symbolizing restricted web data.

Machine Learning Research

Web Data Diminishes: What if online publishers make it harder and more expensive to train models?

For decades, AI developers have treated the web as an open faucet of training data. Now publishers are shutting the tap. Will web data dry up?

Flowchart details GEPA algorithm, featuring candidate filtering and performance improvement loops.

Machine Learning Research

Better Agentic Prompts Automatically: Authors devised GEPA, an algorithm for better prompts to improve agentic systems’ performance

Honing an agent’s prompt can yield better results than fine-tuning the underlying large language model via reinforcement learning.

Graph showing increasing security risks from 9% to 92% as MCP servers rise from 1 to 10.

Machine Learning Research

MCP Poses Security Risks: Experts identify holes in the popular Model Context Protocol for attackers to access data

The ability to easily connect large language models to tools and data sources has made Model Context Protocol popular among developers, but it also opens security holes, research shows.

A performance comparison table highlights Ling-1T's success in reasoning and coding tasks against rivals.

Machine Learning Research

Reasoning Without “Thinking”: All about Ant Group’s Ling-1T, an open, non-reasoning model that outperforms closed competitors

Reasoning models typically learn to undertake a separate process of “thinking” through their output of before they produce final response. Ant Group built a top non-reasoning model that can take similar steps as part of its immediate response.

Image shows MolmoAct system processing "Put the plate in the dishwasher" through spatial reasoning.

Machine Learning Research

Better Spatial Perception for Robots: MolmoAct creates spatial maps for robots to plot their actions before executing text directions

Robot control systems that accept only text input struggle to translate words into motions in space. Researchers developed a system that enables robots to plan spatial paths before they execute text instructions.

Close-up of a violin scroll and pegs, symbolizing precision needed in fine-tuning AI models.

Machine Learning Research

Fine-Tuning Simplified: Thinking Machines’ new Tinker API makes it easier to fine-tune models on many GPUs

The first offering from Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati, aims to simplify — and democratize — the process of fine-tuning AI models.

Graphs compare DeepSeek models showing reduced cost per million tokens with V3.2-Exp over V3.1-Terminus.

Machine Learning Research

DeepSeek Cuts Inference Costs: DeepSeek-V3.2-Exp streamlines processing using a "lightning indexer," boosting efficiency

DeepSeek’s latest large language model can cut inference costs by more than half and processes long contexts dramatically faster relative to its predecessor.

Flowchart of Text-to-LoRA model processes task embeddings into LoRA adapters, showing weights and losses.

Machine Learning Research

LoRA Adapters On Tap: Text-to-LoRA generates task-specific LoRA adapters directly from natural language descriptions

The approach known as LoRA streamlines fine-tuning by training a small adapter that modifies a pretrained model’s weights at inference. Researchers built a model that generates such adapters directly.

Bar chart comparing performance of Qwen3 models against others in diverse tasks, highlighting Qwen3-Max.

Machine Learning Research

Qwen3 Goes Big (and Smaller): Alibaba expands Qwen3 family with a 1 trillion-parameter Max model, open-weights Qwen3-VL, and the Qwen3-Omni voice model

Alibaba rounded out the Qwen3 family with its biggest large language model to date as well as smaller models that process text, images, video, and/or audio.

Comparison table highlighting Claude Sonnet 4.5's top scores in coding and reasoning benchmarks, featuring improved capabilities.

Machine Learning Research

Claude Levels Up: Anthropic launches Claude Sonnet 4.5 and the Claude Agent SDK, and overhauls Claude Code for developers

Anthropic updated its mid-size Claude Sonnet model, making it the first member of the Claude family to reach version 4.5. It also enhanced the Claude Code agentic coding tool with long-desired features.

Image illustrates data flow from raw satellite sources through processing to embeddings for climate tracking.

Machine Learning Research

Earth Modeled in 10-Meter Squares: Google’s AlphaEarth Foundations tracks the whole planet’s climate, land use, potential for disasters, in detail and at scale

Researchers built a model that integrates satellite imagery and other sensor readings across the entire surface of the Earth to reveal patterns of climate, land use, and other features.

Electron microscope image of bacteriophages with distinct hexagonal heads and tails on a gray background.

Machine Learning Research

AI Generates Viral Genomes: Researchers use genomic language models to create custom viruses

Researchers used AI models to create novel viruses from scratch.