SevenTnewS

NLP & ML

Natural language processing, machine learning and deep learning.

6 published articles

Featured3 min read

AI Research

Why GPT-5.5 dominates a benchmark that tests how agents improve themselves

EvoPolicyGym isolates a critical but understudied capability: an agent's ability to refine an executable policy through repeated feedback-constrained edits. The benchmark reveals GPT-5.5 as the strongest performer across 16 environments, and provides trajectory-level diagnostics that expose how different agents allocate budget and convert feedback into tuned parameters.

2026-07-11

5 min read

Memory management

The 12,000-line secret behind Bing's speed: how mimalloc beat the allocator trade-off

Microsoft Research's mimalloc allocator uses thousands of per-page free lists and a clever page-stealing technique to achieve both high concurrency and low memory overhead, as detailed by the RiSE group in a new technical blog.

2026-07-10

5 min read

NLP history

The benchmark that made language models speak: how 2018's glue bet changed ai forever

The GLUE benchmark, launched in 2018, transformed natural language processing by providing a standardized yardstick for language understanding. Its legacy lives on in every modern LLM benchmark, from SuperGLUE to the latest arena-style evaluations that define today's AI race.

2026-07-08

2 min read

Machine Learning Research

Fast-LeWM: Parallel Action-Prefix Prediction Slashes Latent World Model Planning Costs

Researchers introduce Fast-LeWM, a latent world model that accelerates visual planning by predicting future states from action prefixes in parallel. The approach cuts computational costs and error buildup, outperforming prior one-step transition models.

2026-07-07

5 min read

Agent Frameworks

IBM's open-source CUGA agent harness skips the plumbing, goes straight to the prompt

IBM's open-source CUGA framework flips the typical agent development model on its head by handling orchestration, state management, and planning. Developers are left to write only a tool list and a prompt. Over two dozen single-file apps demonstrate the approach, from a movie recommender to a multi-agent lead-generation system, all deployable in governed production without needing a rewrite.

2026-07-07

2 min read

Reinforcement Learning Research

OPID gives language agents a reward signal dense enough to ditch external memory

OPID extracts hierarchical skill supervision from completed on-policy trajectories, providing dense token-level guidance for language agent training without external memory. Experiments on ALFWorld, WebShop, and Search-based QA show improved performance and sample efficiency over outcome-only RL and existing skill-distillation methods.

2026-07-06