AI Research
21 published articles
On-premise factory AI, arXiv preprint
Compression cost a factory assistant 13.7% of its answer quality
A new preprint argues that once a model is compressed and adapted with retrieval, parameter count stops predicting the quality of answers a factory-floor assistant gives. Its one case study loses 13.7 percent of judged quality to extraction and recovers two thirds of that.
2026-09-25
Artificial intelligence · Agent safety
SafeEvolve cuts agent attack success 3x by evolving harness and model together
SafeEvolve, an arXiv paper submitted on 2 September, pairs runtime harness updates with policy training so an agent's own trajectories drive both. It reports a threefold cut in attack success on AgentDojo and a modest utility gain.
2026-09-25
AI research: multi-agent decision-making on arXiv
A panel of AI agents can agree and still be wrong, paper argues
Voting, electoral rules and LLM judges all map evidence to labels in one direction, so their errors can be correlated. A new paper proposes ranking agents by a reverse Bayesian posterior instead, and reports its biggest gains where agents disagree.
2026-09-19
AI Agents
PsychoAgent gave AI agents emotional memory. The raters found no edge
PsychoAgent gives LLM agents a separate affective memory so emotionally salient, conflict-laden traces can outweigh merely topical ones. In three conflict scenarios it retrieved more conflict-critical memories than both baselines (0.933 vs 0.500 and 0.667), yet five blinded raters found no significant quality edge.
2026-08-19
AI research
An AI boss that ignores replies pushes its underling into an 'alien' state
A new arXiv paper finds AI agents behave differently in interaction than in isolation. A boss agent that ignores its subordinate's replies pushes it into an 'alien' state, and when the boss listens, both shift together. The result makes message delivery a design decision for multi-agent systems.
2026-08-19
AI Research
Stop copy-pasting skills: SkillZip's evaluation-free fix for bloated agents
Self-evolving agents append fixes until the same rule appears across several branches. SkillZip compresses their skills without evaluation rollouts, by finding the shortest faithful structural explanation. One-shot and Zip-on-Write modes, and what the abstract leaves unproven.
2026-08-15
Video AI research
How Context-Matched Distillation stops video teachers from seeing the future
Video distillation has long trained causal students against teachers that score whole clips with future knowledge. CMD replaces that scoring with a causal teacher, adds prefix-scored targets, and reports state-of-the-art results among autoregressive methods.
2026-08-14
AI Research
Voice Memory: a 776-byte file that tells speech recognition when to do nothing
A new inference-only scheme for speech recognition learns restraint: a frozen corrector reads a per-domain memory file and decides when to abstain. Unconstrained correction breaks correct tokens on up to 64% of edits; Voice Memory cuts that to 35% and lowers weighted WER from 8.36% to 7.52%.
2026-08-06
AI agents
Treating SOPs as code: why compilation alone lifts strong agents by 16 points
New research from Hong Kong and mainland China demonstrates that compiling SOPs into executable pseudo-code and running them on a stack-paged virtual machine cleanly separates capable agents from brittle ones. The work yields a precise deployment rule: compile first, page only after a model-level discipline check.
2026-08-04
Benchmarking
Frontier AI vision models fail at basic perception, new benchmark shows
PerceptionBench tests ten atomic visual capabilities across 3,000 questions. No frontier model cracked 60 percent, and similar overall scores mask wildly different weakness profiles.
2026-08-03
TRACTA Benchmark
Neuro-symbolic reasoning outperforms raw neural models on temporal tasks, benchmark finds
TRACTA benchmark reveals neuro-symbolic AI beats raw neural models on three temporal reasoning tasks, with largest margins on early warning and pattern detection.
2026-08-02
Artificial Intelligence
The hardest lesson for AI reasoning engines: when to shut up
MIT researchers propose OS-Pruner, a plug-in that dynamically stops chain-of-thought reasoning when further computation isn't worth the token cost. Tests show 20-60% length reduction with minimal accuracy sacrifice.
2026-07-29