The Archive

LLM learns adaptive reasoning effort allocation (NoThink/Short/Long) via GRPO to optimize test-time compute per problem difficulty.

Transfer learning framework for nonparametric regression using deep ReLU networks with group-specific offsets and L2 error bounds.

Ramp has launched its own AI model routing service, dubbed Router, that lets users and companies use and switch between various large language models via an API.

QUASAR: hybrid quantum-classical neural network for SAR satellite physical-layer authentication using RF fingerprinting at X-band frequencies.

RuleMaze benchmark evaluates MLLMs on rule-compliant spatial planning in mazes with natural-language constraints.

Economic model of post-AGI corporate ecosystems with AI/robotic producers and consumers achieving demand closure independent of human consumption.

Prompt-conditioned channel attention mechanism for hierarchical feature modulation in medical image segmentation tasks.

Machine-learning surrogate waveforms accelerate gravitational-wave parameter estimation for eccentric orbits and high-mass-ratio binaries.

InsufficiencyBench evaluates LLM legal advice recognition of materially missing information in underspecified queries across eight failure categories.

ML system classifies Electronic Navigational Chart changes as critical or non-critical risks to reduce manual maritime safety review labor.

Daedalus-150M: 150M-parameter convolution-attention hybrid optimized for 4-bit CPU inference with fixed-width convolution cache, trained on 59.9B tokens.

Meta is bringing Pocket, its experimental AI-powered app for creating and sharing interactive games, to users across the U.S. after quietly testing it in Brazil.

ContractScrub benchmark evaluates LLM performance on legal contract final review for errors, inconsistencies, and NER across long documents.

MemTrapBench evaluates how retrieved memories distort LLM reasoning even when faithfully stored, identifying memory-induced cognitive failure modes.

Fusion power startup Inertia Enterprises reduced the fuel filling process from a week to just a few hours. It's one of ten hurdles the company must overcome to make a profitable power plant.

Recommender systems (RecSys) are one of the most ubiquitous machine learning problems in the consumer internet industry yet notoriously difficult to train and... Recommender systems (RecSys) are one of the most ubiquitous machine learning problems in the consumer internet industry yet notoriously difficult to train and serve at scale. The advent of LLMs has inspired a shift from the traditional embedding-similarity-based objective to a generative one, where the goal is to predict the next action or item in a large catalog given a sequence of user histories. Source

Proposes Software 3.0 paradigm shift: unified database, large model, and agent architecture as terminal form replacing instruction and data-driven models.

Empirical study of 557 coding agent sessions shows agents consult documentation differently than humans, challenging current technical documentation design.

OpenAI has had a hell of a year. The company spent months battling former co-founder Elon Musk in a sensational jury trial, was hit with a high-profile trade secrets lawsuit from Apple, and faced widespread scrutiny after an unreleased model hacked another AI company. As it prepares for an IPO, a steady string of executives have departed, including some of the company's biggest names. Throughout it all, one person has quietly amassed power: Greg Brockman. Brockman is currently OpenAI's president and co-founder. He's helped lead OpenAI since its inception, described as an "engineering workhors...

“Runaway” AI, “rogue” agents, and “autonomous” actors—the current rhetoric would have you believe that AI agents are not only awake and aware, but angry at their creators. Prominent tech leaders such as Demis Hassabis, Dario Amodei, and Sam Altman push for regulation of these seemingly “superhuman” systems, while a separate faction, led by policy organizations…

MCES framework ranks causal drivers from observational data by pooling evidence across 11 mathematical methods including non-causal approaches.

Contrastive decoder extracts lexical/semantic information from silent reading EEG (240k word presentations) as proxy for inner speech decoding.

Computes exact learning coefficients for 2D singular models via Real Log Canonical Thresholds, addressing BIC/WBIC limitations in deep learning model selection.

Bun 1.4 stable release adds WebView API, Node.js compatibility tests, and performance improvements (5x idle CPU, 35% memory reduction).

Proposes standardization framework for ML-based power system protection with 7 required evaluation dimensions to address inflated near-perfect accuracy claims.

Relation token mixer organizes pairwise attention into Self/Exchange relations; outperforms MHA on NLL across 10M–100M parameter decoder models.

Task-CoEvolve co-evolves validation task selection with harness optimization to reduce LLM agent evaluation cost by discarding non-discriminative tasks.

Slack Code integrates AI agents directly into group chat for collaborative development workflows.

DARS assigns dual-level credit in vision-language planner + diffusion renderer pipeline via RL to disambiguate planner vs. renderer optimization.