Improving understanding with language
Source: MIT News – Artificial intelligence When she was a child, MIT senior Olivia Honeycutt would spend summers...
Moonshot AI Open-Sources FlashKDA: CUTLASS Kernels for Kimi Delta Attention with Variable-Length Batching and H20 Benchmarks
Source: MarkTechPost The team behind Kimi.ai (Moonshot AI) just made a significant contribution to the open-source AI infrastructure...
Microsoft Research’s World-R1 Uses Flow-GRPO and 3D-Aware Rewards to Inject Geometric Consistency Into Wan 2.1 Without Architectural Changes
Source: MarkTechPost Video foundation models can paint a beautiful frame. They are still notoriously bad at remembering it....
A Coding Implementation on Pyright Type Checking Covering Generics, Protocols, Strict Mode, Type Narrowing, and Modern Python Typing
Source: MarkTechPost In this tutorial, we explore Pyright, Microsoft’s high-performance static type checker for Python, and walk through...
IBM Releases Two Granite Speech 4.1 2B Models: Autoregressive ASR with Translation and Non-Autoregressive Editing for Fast Inference
Source: MarkTechPost IBM released two new open speech recognition models— Granite Speech 4.1 2B and Granite Speech 4.1...
Cursor Introduces a TypeScript SDK for Building Programmatic Coding Agents With Sandboxed Cloud VMs, Subagents, Hooks, and Token-Based Pricing
Source: MarkTechPost Cursor, the AI-powered code editor, is opening up the core technology behind its coding agents to...
Making the case for curiosity-driven science
Source: MIT News – Artificial intelligence “The thing that really struck me when I came to MIT and...
Solving the “Whac-a-mole dilemma”: A smarter way to debias AI vision models
Source: MIT News – Artificial intelligence In today’s hospitals and clinics, a dermatologist may use an artificial intelligence...
Top 10 KV Cache Compression Techniques for LLM Inference: Reducing Memory Overhead Across Eviction, Quantization, and Low-Rank Methods
Source: MarkTechPost As large language models scale to longer context windows and serve more concurrent users, the key-value...
Qwen Team Releases FlashQLA: a High-Performance Linear Attention Kernel Library That Achieves Up to 3× Speedup on NVIDIA Hopper GPUs
Source: MarkTechPost The race to make large language models faster and cheaper to run has largely been fought...