Knowledge catalogue
Will Douglas Heaven / MIT Technology Review: An interview with OpenAI Chief Research Officer Mark Chen on the Hugging Face incident, slowing AI development, shifting 5%-10% of compute to safety, and m
'AREX-2 is a 27B-parameter long-horizon agent model from the Beijing Academy of Artificial Intelligence (BAAI). It learns to improve a solution over multiple test-time rounds: propose, measure, reflec
arXiv:2601.04094v4 Announce Type: cross Abstract: Effective regulation of AI is a defining policy challenge, driven by their integration into all aspects of society. To remain responsive to their rapi
arXiv:2609.36359v1 Announce Type: new Abstract: Graph-based approximate nearest neighbor search (ANNS) is widely used for large-scale semantic search. Its indices are constructed primarily based on ge
arXiv:2602.11678v2 Announce Type: replace Abstract: Multimodal Large Language Models (MLLMs) have shown remarkable progress in visual understanding, yet they suffer from a critical limitation: structu
arXiv:2609.35841v1 Announce Type: cross Abstract: Mutation testing evaluates test-suite adequacy by injecting synthetic faults into program code. However, traditional rule-based tools often generate l
arXiv:2609.36776v1 Announce Type: new Abstract: Emotional Video Captioning aims to generate factually accurate and emotionally empathetic descriptions. While recent methods have recognized the importa
arXiv:2609.36239v1 Announce Type: new Abstract: Large language models (LLMs) can predict human brain activity across a variety of brain regions during natural language comprehension. Typically, howeve
Tasks where an MLLM doesn't write text but only makes a decision, which started with Jev, seem to be getting popular. (This is sometimes called 'typed decision'.) I thought it would be fun to use insi
arXiv:2602.20731v2 Announce Type: replace-cross Abstract: Discrete image tokenizers provide a sequential interface for vision and multimodal models, but are typically optimized for reconstruction or c
arXiv:2609.37661v1 Announce Type: new Abstract: Graph-based retrieval-augmented generation supports multi-hop retrieval by organizing corpus information into graphs. However, existing relation-free gr
arXiv:2609.37841v1 Announce Type: new Abstract: Masked generative models offer parallel token prediction, but accurate parallel sampling must account for dependencies among tokens. When dependencies a
arXiv:2609.36570v1 Announce Type: cross Abstract: Indirect prompt injection makes an LLM agent treat untrusted retrieved text as instructions. We present CounterSteer, an inference-time defense that s
arXiv:2609.36987v1 Announce Type: new Abstract: Graph databases are increasingly queried through natural language, yet every existing benchmark evaluates isolated single-turn queries rather than the m
arXiv:2609.36344v1 Announce Type: new Abstract: Deep-research agents conduct long-horizon investigations through iterative search, evidence evaluation, belief revision, and synthesis. However, they ma
arXiv:2609.36734v1 Announce Type: new Abstract: Knowledge Distillation (KD) trains a smaller-capacity student model to imitate a larger-capacity teacher model by matching output distributions, implici
arXiv:2609.38025v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on its own generated responses using dense, token-level supervision from a stronger teacher. Vanilla OPD
arXiv:2609.36845v1 Announce Type: cross Abstract: Uncrewed aerial vehicle base stations (UAV-BSs) are expected to cover traffic demand that shifts across space and time, yet most repositioning schemes
arXiv:2609.36944v1 Announce Type: new Abstract: Group-relative policy optimization relies on reward-derived advantages and sequence-level likelihood weights, both of which can be sensitive to localize
arXiv:2609.36116v1 Announce Type: new Abstract: We study how to build more capable general-purpose agents by extending large language models (LLMs) with native typed decision-making. We introduce Dyad
arXiv:2609.36802v1 Announce Type: new Abstract: A key strength of Proximal Policy Optimization (PPO) is its learned critic, which uses historical trajectories collected during reinforcement learning t
arXiv:2609.36365v1 Announce Type: new Abstract: Can interaction formats and textual scaffolds help large language model (LLM) agents make better decisions, and do better decisions come with better exp
arXiv:2609.37686v1 Announce Type: new Abstract: Autonomous agents have made rapid progress in general-purpose computer use, but reliable automation of professional industrial engineering remains out o
arXiv:2609.37647v1 Announce Type: cross Abstract: Jev is a commercial System One model from TypeSafe AI that does not generate text: given a state and typed questions, it returns a choice from fixed o
arXiv:2602.18998v2 Announce Type: replace Abstract: LLM agents are increasingly expected to operate as general-purpose systems that resolve real-world user requests, yet their dynamic scaling behavior
arXiv:2609.36187v1 Announce Type: new Abstract: Symbolic Regression (SR) is a data-driven method for scientific discovery which searches for interpretable analytical relationships within data. Recentl
arXiv:2609.37655v1 Announce Type: new Abstract: Advancing spatial intelligence in Multimodal Large Language Models (MLLMs) is bottlenecked by the scarcity of complex, scalable 3D question-answer (QA)
arXiv:2609.36172v1 Announce Type: new Abstract: Figuring out how objects relate to each other, like whether they touch, overlap, stay completely separate or one sits inside another, matters a lot in f
arXiv:2609.35791v1 Announce Type: new Abstract: Natural turn-taking in full-duplex voice interaction requires determining from partial speech whether a pause reflects hesitation or a completed convers
https://preview.redd.it/xs01ds930nsh1.png?width=4800&format=png&auto=webp&s=939762bcf3fbd93e137d4f97e927dda38aaf30fd Back in July I shared LokalBot here. It's a free, open-source Mac app that records
arXiv:2609.36617v1 Announce Type: new Abstract: We study the ability of LLMs to generate edit-inducing questions whose answer will improve a paper draft. On a dataset of paired submission and camera-r
arXiv:2609.37190v1 Announce Type: new Abstract: Multi-turn visual search agents answer questions about high-resolution images by iteratively deciding where to look. Reinforcement learning for these ag
arXiv:2609.38006v1 Announce Type: new Abstract: Running a language model locally offers advantages in privacy, latency, and cost, but local hardware fits only small models, which are less capable than
arXiv:2609.36289v1 Announce Type: cross Abstract: Prompts are the primary mechanism for directing the behavior of large language models (LLMs). Yet the internal structure and causal hierarchy of promp
I've been closely following the rise of the Strata inference engine and as someone with 28GB VRAM and 32GB RAM I'm itching to buy 32GB more RAM just to use Flash Next. But of course before I make such
arXiv:2609.36265v1 Announce Type: cross Abstract: Large language models can learn abstract rules from just a few in-context examples, but how their internal mechanisms activate as examples accumulate
Independent numbers from @ValsAI. Fewer tokens per turn means lower cost and faster agent loops on every call. Try Ember-1 on Fireworks Serverless: https://fireworks.ai/models/fireworks/ember-1 Firewo
arXiv:2609.38065v1 Announce Type: cross Abstract: Training agents to follow arbitrary instructions is an important goal of multi-task reinforcement learning (RL). Linear temporal logic (LTL) provides
arXiv:2609.37988v1 Announce Type: new Abstract: As the context size of text processed with an LLM grows, the size of KV caches can outstrip the memory allocated for the original model weights. This im
arXiv:2609.36139v1 Announce Type: cross Abstract: As large language models are deployed in increasingly autonomous long-horizon tasks, manually auditing and verifying the actions, artifacts, and outpu
arXiv:2609.37017v1 Announce Type: new Abstract: LLM-based multi-agent systems (MAS) increasingly use latent collaboration to avoid the information loss and repeated encoding-decoding overhead of natur
arXiv:2609.36371v1 Announce Type: cross Abstract: Test-time scaling improves software engineering agents by generating multiple candidate trajectories and selecting the best one. Verifying and selecti
arXiv:2609.36932v1 Announce Type: new Abstract: Recently, Group Relative Policy Optimization (GRPO) and its variants have been developed for policy optimization and demonstrated notable performance ga
arXiv:2609.37868v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) methods such as GRPO rely on successful self-generated trajectories, but finite rollout budgets
arXiv:2609.35817v1 Announce Type: cross Abstract: Although uniform diffusion language models (UDLMs) represent a promising diffusion paradigm, scaling them remains challenging. We identify the core ob
arXiv:2609.31857v2 Announce Type: replace Abstract: Validating an LLM-as-a-judge requires estimating its agreement with humans, yet annotation budgets rarely allow every item to be multiply labeled. W
arXiv:2609.34187v2 Announce Type: replace-cross Abstract: The strong version of the stochastic parrot argument claims that, although large language models (LLMs) may exceed rote regurgitation, they ca
arXiv:2609.37408v1 Announce Type: new Abstract: Narrative extraction allows us to identify online hate narratives, supporting the construction of rigorous detection systems. Existing computational app
arXiv:2609.36059v1 Announce Type: cross Abstract: Long-term memory lets an LLM assistant use a history it can no longer reread, and most memory systems build it by rewriting conversations into facts,
arXiv:2609.36301v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) layers are central to frontier language models, and recent architectures push toward more and smaller experts. In this regime
arXiv:2609.38079v1 Announce Type: new Abstract: Training a model to generate visual content can encourage it to learn rich perceptual capabilities related to geometry, spatial relationships, and objec
I just bought the GPUs after deliberating and debating for over two years. With costs not coming down any time soon and me just wanting to get this massive todo-box ticked, I decided to just YOLO it;
arXiv:2609.37809v1 Announce Type: cross Abstract: Predicting canopy height from medium-resolution satellite imagery is a common and scalable approach for assessing the condition of the world's forests
arXiv:2609.36543v1 Announce Type: new Abstract: Partial observability remains a fundamental challenge in robotic navigation, where limited sensor coverage and occlusions leave large portions of the en
arXiv:2609.38149v1 Announce Type: new Abstract: Transformer language models (LMs) are feed-forward: deep-layer representations are never fed back to shallower layers, and the only pathway for informat
arXiv:2609.35816v1 Announce Type: cross Abstract: Large language model search agents are often trained with synthetic questions whose difficulty is increased through larger evidence graphs, additional
arXiv:2609.37681v1 Announce Type: new Abstract: Long-horizon robotic tasks are vulnerable to unexpected environmental changes that can render planned actions ineffective or unsafe. To address this, ro
arXiv:2609.35928v1 Announce Type: cross Abstract: Multi-agent LLM systems increasingly mix models from several providers, yet exposing each agent's underlying model identity to its peers significantly
arXiv:2609.36760v1 Announce Type: cross Abstract: Multi-Head Latent Attention (MLA) enables expressive multi-head attention with compact caches for its content and decoupled RoPE paths, yet cache memo
arXiv:2609.37734v1 Announce Type: new Abstract: Research in modular robotics has produced capable approaches allowing a robot's morphology to change online, with recent efforts also developing approac