Knowledge catalogue
I find it fascinating that OpenAI's Global Affairs team will denounce AI doomers for being too negative and then, simultaneously, advocate for an Illinois bill that would shield them from liability if
I spent some time trying to distill all the complex factors impacting open models -- economics, capabilities, distribution, policy, etc. -- into a clear list of beliefs. Here they are in full. 1. It’s
arXiv:2505.19134v2 Announce Type: replace-cross Abstract: Human-annotated data plays a vital role in training large language models (LLMs), such as supervised fine-tuning and human preference alignmen
arXiv:2604.12340v1 Announce Type: cross Abstract: We decompose the Kullback--Leibler generalization error (GE) -- the expected KL divergence from the data distribution to the trained model -- of unsup
arXiv:2604.12721v1 Announce Type: new Abstract: Clinical case formulation organizes patient symptoms and psychosocial factors into causal models, often using the 5P framework. However, constructing su
arXiv:2604.12303v1 Announce Type: new Abstract: Batch active learning (BAL) is a crucial technique for reducing labeling costs and improving data efficiency in training large-scale deep learning model
arXiv:2604.12413v1 Announce Type: cross Abstract: Dynamic soaring enables sustained flight by extracting energy from wind shear, yet it is commonly understood as a cycle-level maneuver that assumes st
arXiv:2604.13015v1 Announce Type: new Abstract: Humanoid robots promise general-purpose assistance, yet real-world humanoid loco-manipulation remains challenging because it requires whole-body stabili
arXiv:2604.13010v1 Announce Type: cross Abstract: On-policy distillation (OPD) has emerged as an efficient post-training paradigm for large language models. However, standard OPD requires a live teach
arXiv:2604.12286v1 Announce Type: new Abstract: Live Photo captures both a high-quality key photo and a short video clip to preserve the precious dynamics around the captured moment. While users may c
arXiv:2603.19042v4 Announce Type: replace Abstract: The integration of artificial intelligence (AI) technologies into judicial decision-making, particularly in pretrial, sentencing, and parole context
arXiv:2604.12479v1 Announce Type: new Abstract: Recent advances in large language models (LLMs) have significantly improved the alignment of models with general human preferences. However, a major cha
arXiv:2604.12277v1 Announce Type: new Abstract: Pretrained language models often rely on superficial features that appear predictive during training yet fail to generalize at test time, a phenomenon k
arXiv:2604.12537v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) have achieved remarkable progress in multimodal understanding, yet their positional encoding mechanisms remain suboptima
arXiv:2604.12237v1 Announce Type: cross Abstract: In drug discovery, molecular optimization aims to iteratively refine a lead compound to improve molecular properties while preserving structural simil
arXiv:2508.17403v3 Announce Type: replace Abstract: A community of researchers appears to think that a machine can be surprised and have introduced various surprise measures, principally the Shannon S
arXiv:2601.06794v2 Announce Type: replace Abstract: Critique-guided reinforcement learning (RL) has emerged as a powerful paradigm for training LLM agents by augmenting sparse outcome rewards with nat
arXiv:2604.05164v2 Announce Type: replace-cross Abstract: As LLM reasoning performance plateau, improving inference-time compute efficiency is crucial to mitigate overthinking and long thinking traces
arXiv:2604.11994v1 Announce Type: new Abstract: We study offline-online reinforcement learning in linear mixture Markov decision processes (MDPs) under environment shift. In the offline phase, data ar
arXiv:2604.12852v1 Announce Type: new Abstract: Collaborative transport requires robots to infer partner intent through physical interaction while maintaining stable loco-manipulation. This becomes pa
arXiv:2507.06448v5 Announce Type: replace Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has proven to be a highly effective strategy for endowing Large Language Models (LLMs) with ro
arXiv:2604.12113v1 Announce Type: cross Abstract: Visual Foundation Models (VFMs) such as the Segment Anything Model (SAM) have significantly advanced broad use of image segmentation. However, SAM and
Prediction from a year ago about AI backlash that looks to be on track: Public backlash against AI will so be strong by 2028 that anti-AI sentiment will probably be a major factor in the 2028 US Presi
arXiv:2511.17097v2 Announce Type: replace Abstract: Vision-Language Navigation requires agents to act coherently over long horizons by understanding not only local visual context but also how far they
arXiv:2604.12160v1 Announce Type: new Abstract: Reasoning post-training with reinforcement learning from verifiable rewards (RLVR) is typically studied in centralized settings, yet many realistic appl
arXiv:2604.12175v1 Announce Type: new Abstract: Recent advances in image editing have heightened the need for reliable Image Editing Quality Assessment (IEQA). Unlike traditional methods, IEQA require
arXiv:2604.12582v1 Announce Type: new Abstract: Recent Video Large Language Models (Video-LLMs) have demonstrated strong capability in video understanding, yet they still suffer from hallucinations. E
arXiv:2604.13016v1 Announce Type: cross Abstract: On-policy distillation (OPD) has become a core technique in the post-training of large language models, yet its training dynamics remain poorly unders
arXiv:2511.09803v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) improves factuality but retrieving for every query often hurts quality while inflating tokens and latency. We p
arXiv:2512.04585v3 Announce Type: replace Abstract: Segment Anything Model 3 (SAM3) advances open-vocabulary segmentation through promptable concept segmentation, enabling users to segment all instanc
arXiv:2604.12350v1 Announce Type: cross Abstract: Molecular property optimization is central to drug discovery, yet many deep learning methods rely on black-box scoring and offer limited control over
arXiv:2602.05791v2 Announce Type: replace Abstract: Learning-based whole-body controllers have become a key driver for humanoid robots, yet most existing approaches require robot-specific training. In
arXiv:2604.11835v1 Announce Type: cross Abstract: Machine learning for tabular data remains constrained by poor schema generalization, a challenge rooted in the lack of semantic understanding of struc
arXiv:2604.12002v1 Announce Type: new Abstract: Current post-training methods in verifiable settings fall into two categories. Reinforcement learning (RLVR) relies on binary rewards, which are broadly
arXiv:2507.08977v4 Announce Type: replace-cross Abstract: Scientific modeling faces a tradeoff between the interpretability of mechanistic theory and the predictive power of machine learning. While ex
So true. “Being right too soon is socially unacceptable” I remember reading this from Heinlein as a young man and it took experience for me to fully understand it. People and institutions will persist
arXiv:2604.12617v1 Announce Type: cross Abstract: The post-training pipeline for diffusion models currently has two stages: supervised fine-tuning (SFT) on curated data and reinforcement learning (RL)
arXiv:2510.20093v2 Announce Type: replace-cross Abstract: Although recent advancements in diffusion models have significantly enriched the quality of generated images, challenges remain in synthesizin
arXiv:2604.12935v1 Announce Type: new Abstract: Efficiently merging several models fine-tuned for different tasks, but stemming from the same pretrained base model, is of great practical interest. Des
arXiv:2604.12770v1 Announce Type: new Abstract: Editing human-written text has become a standard use case of large language models (LLMs), for example, to make one's arguments more appropriate for a d
The front door to the internet hasn't changed -- but the person going through that door has changed from a human browsing 5 blue links, to an agent browsing a similar index in vastly different ways. A
arXiv:2604.12816v1 Announce Type: new Abstract: Implicit biases in both humans and large language models (LLMs) pose significant societal risks. Dual process theories propose that biases arise primari
arXiv:2510.09087v2 Announce Type: replace Abstract: Large language model (LLM) agents have shown remarkable progress in social deduction games (SDGs). However, existing approaches primarily focus on i
arXiv:2509.25758v2 Announce Type: replace Abstract: The remarkable capabilities of modern large reasoning models are largely unlocked through post-training techniques such as supervised fine-tuning (S
arXiv:2604.12736v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO) has significantly advanced the reasoning ability of large language models (LLMs), particularly in their mathem
arXiv:2410.03000v3 Announce Type: replace Abstract: Existing certified training methods can only train models to be robust against a certain perturbation type (e.g. l_infty or l_2). However, an l_inft
arXiv:2604.12133v1 Announce Type: new Abstract: Historical approaches to Table Representation Learning (TRL) have largely adopted the sequential paradigms of Natural Language Processing (NLP). We argu
arXiv:2603.05295v3 Announce Type: replace Abstract: We introduce WebChain, the largest open-source dataset of human-annotated trajectories on real-world websites, designed to accelerate reproducible r
What happens when you systematically oversell the value of your product for years, while pretty much screwing society along the way? Eventually your customers figure it out. Surprisingly, people see A
arXiv:2604.12509v1 Announce Type: cross Abstract: Mobile Manipulation (MoMa) of articulated objects, such as opening doors, drawers, and cupboards, demands simultaneous, whole-body coordination betwee
arXiv:2604.12025v1 Announce Type: new Abstract: The Semantic Web standardizes concept meaning for humans and machines, enabling machine-operable content and consistent interpretation that improves adv
arXiv:2604.13001v1 Announce Type: new Abstract: The acquisition of high-quality, action-aligned demonstration data remains a fundamental bottleneck in scaling foundation models for dexterous robot man
arXiv:2604.09639v1 Announce Type: new Abstract: Artistic style transfer is well studied for images and videos, but extending it to multi-view 3D scenes remains difficult because stylization can disrup
arXiv:2604.09676v1 Announce Type: cross Abstract: Reinforcement learning (RL) has become a key approach for enhancing reasoning in large language models (LLMs), yet scalable training is often hindered
arXiv:2604.10252v1 Announce Type: new Abstract: Reinforcement learning agent-based simulation (RL-ABS) has become an important tool for electricity market mechanism analysis and evaluation. In the mod
arXiv:2601.16399v5 Announce Type: replace Abstract: We study a structured bi-level optimization problem where the upper-level objective is a smooth function and the lower-level problem is policy optim
arXiv:2604.05142v2 Announce Type: replace Abstract: As artificial intelligence systems (AIs) become increasingly produced by recursive self-improvement, a form of evolution may emerge, with the traits
arXiv:2407.10953v5 Announce Type: replace Abstract: The Mutual Reinforcement Effect (MRE) describes a phenomenon in information extraction where word-level and sentence-level tasks can mutually improv
Byard Duncan / Bloomberg: A profile of BusPatrol, whose AI-powered cameras on 35K+ school buses in 24 US states record vehicles passing illegally, claiming they help reduce violations — BusPatrol says
arXiv:2604.11125v1 Announce Type: new Abstract: India generates vast biomedical data through postgraduate research, government hospital services and audits, government schemes, private hospitals and t