Knowledge catalogue
arXiv:2605.20246v1 Announce Type: new Abstract: Recently, vision-language model (VLM) agents have shown promising progress in open-world tasks, where successful task completion often requires multiple
Here's how the corruption works: Thursday: RJ Reynolds donates $5M to Trump Saturday: Trump invites RJR execs to Mar a Lago; execs ask to loosen regs on flavored vapes; Trump calls up RFK Jr. and tell
arXiv:2605.21104v1 Announce Type: new Abstract: Sparsifying transformers remains a fundamental challenge, as standard optimizers fail to simultaneously encourage sparsity and maintain training stabili
I don't think anyone has a good intuitive sense about what this means, and that failure of imagination is a generally bad thing for planning, investment, and policy. I also don't have an easy solution
I was happy to join the @ScienceBoard_UN podcast to discuss deception among frontier AI models and the need to manage its potential global impacts and risks. 🎙️What happens when AI learns to lie? Scie
if you own a tesla, you are the product Astonishing potential, especially if one takes all the galaxies into account and factors them into the calculation. 😃 And this comes from a man who removed Auto
arXiv:2410.23212v2 Announce Type: replace-cross Abstract: In graph-based data analysis, k-nearest neighbor (kNN) graphs are widely used due to their adaptivity to local data densities. Allowing weight
arXiv:2603.22430v2 Announce Type: replace Abstract: Offline Reinforcement Learning (RL) learns optimal policies from fixed datasets, training a policy once and deploying it at inference time without f
arXiv:2605.20258v1 Announce Type: new Abstract: Contextual Integrity (CI) defines privacy not merely as keeping information hidden, but as governing information flows according to the norms of a given
It’s crazy that OpenAI and xAI were built partly as ways to mitigate @demishassabis’s power and how both are more reckless and less focused on the good part, viz AI for science. And then Anthropic was
arXiv:2605.21362v1 Announce Type: new Abstract: Jailbreak attacks expose a persistent gap between the intended safety behavior of aligned large language models and their behavior under adversarial pro
arXiv:2605.20391v1 Announce Type: cross Abstract: Traditional anomaly detection marks events when measured signals cross predefined thresholds. This captures the moment of transition but not the struc
arXiv:2605.21330v1 Announce Type: new Abstract: In-hand object manipulation is a fundamental yet challenging capability for dexterous robots. Despite significant progress in dexterous manipulation, ex
arXiv:2605.20780v1 Announce Type: cross Abstract: Physics-informed diffusion models typically enforce PDE constraints only on final outputs, leaving intermediate representations unconstrained and pron
arXiv:2510.09060v2 Announce Type: replace-cross Abstract: Flow-based text-to-image models follow deterministic trajectories, making it costly to explore diverse modes under limited sampling budgets. E
arXiv:2605.21123v1 Announce Type: new Abstract: Direct Preference Optimization (DPO) is successful for alignment in LLMs but still faces challenges in text-to-image generation. Existing studies are co
arXiv:2605.06139v2 Announce Type: replace Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a standard approach for large language models (LLMs) post-training to incentivize r
arXiv:2605.20449v1 Announce Type: new Abstract: Can language-pretrained transformers become effective time-series forecasters, and why? In this paper, we show that cross-modal transfer arises because
longer comment on yesterday’s stories the nuance between yesterday’s stories that X influencers won’t tell you: https://open.substack.com/pub/garymarcus/p/checking-the-math-behind-openai-and?r=8tdk6&u
arXiv:2605.21446v1 Announce Type: new Abstract: Interpretable autonomous driving planners depend not only on generating explanations, but also on those explanations remaining reliable under real-world
arXiv:2605.20577v1 Announce Type: cross Abstract: Riichi Mahjong is a multi-player, imperfect-information game characterized by stochasticity and high-dimensional state spaces. These attributes presen
arXiv:2605.20607v1 Announce Type: cross Abstract: EASA's learning-assurance guidance requires data-driven aviation systems to build and monitor their own situation representation, yet for neural netwo
arXiv:2605.21458v1 Announce Type: cross Abstract: Suppose a planner has a pre-trained simulator of a sequential decision problem and the option to run real experiments in the field. The simulator is c
arXiv:2605.21455v1 Announce Type: new Abstract: Statistical decision algorithms are increasingly deployed in domains where ground-truth labels are hard to obtain, such as hiring, university admissions
arXiv:2605.20894v1 Announce Type: new Abstract: Mobile imitation learning on portable demonstration interfaces faces two coupled bottlenecks: locomotion-contaminated action labels and inference-induce
arXiv:2605.20255v1 Announce Type: new Abstract: Simulation-based testing of self-driving cars (SDCs) typically relies on scripted or simplified pedestrian models that do not capture the heterogeneity
arXiv:2605.20271v1 Announce Type: cross Abstract: We develop a rigorous statistical theory of multi-head attention (MHA) as an ensemble of Nadaraya-Watson (NW) kernel regression estimators. Building o
arXiv:2605.20865v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) plays a pivotal role in improving the reasoning ability of large language models. However, widely
arXiv:2605.21059v1 Announce Type: new Abstract: Despite the impressive results achieved by multimodal large language models (MLLMs), their training typically relies on jointly curated multimodal data,
arXiv:2605.20209v1 Announce Type: cross Abstract: Achieving precise, versatile whole-body character control in physics-based animation remains challenging. Recent diffusion-based policies generate ric
arXiv:2605.20302v1 Announce Type: cross Abstract: Supervised classification has a theoretical optimum, Neural Collapse (NC), yet neither of its two dominant paradigms reaches it in practice. Cross ent
nice measure of how insane SpaceX’s claims are I don't offer investment advice, but when I see that SpaceX claims that their total addressable market is the size of the *complete* US economy (with eve
nor do we know how the (new) model works nor how it does on anything else nor how it was trained. scientists wait for facts; cheerleaders (over and over) rush to judgments that have often been wrong.
note that @jeffBezos is not betting on chatbots, but rather focusing on domain-specific tools for engineering, which is also something I would encourage. the financial story for what he is doing may b
arXiv:2605.21343v1 Announce Type: new Abstract: Recent layout-to-image models have achieved remarkable progress in spatial controllability. However, they still struggle with inter-object occlusion. Wh
OpenAI reported a negative 122% operating margin in Q1, indicating the company spent significantly more on operations than it generated in revenue, even after excluding stock-based compensation and ot
our company, which lost money last quarter, is going to take a big slice of 28.5 trillion market that we project will exist, primarily in a field in which we have almost no current presence, which is
arXiv:2605.20194v1 Announce Type: new Abstract: Large language models (LLMs) have been increasingly used to analyze text. However, they are often plagued with contextual reasoning limitations when ana
arXiv:2605.20640v1 Announce Type: new Abstract: Text-to-image diffusion models often face a severe trilemma in human portrait generation: text-image alignment, photorealism, and human-perceived aesthe
arXiv:2605.21414v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have shown strong potential for general-purpose robotic manipulation by leveraging large pretrained vision-languag
arXiv:2605.21422v1 Announce Type: new Abstract: As LLMs continue to scale, improving training efficiency increasingly depends on using data more effectively. Data selection addresses this problem by a
arXiv:2605.21225v1 Announce Type: new Abstract: We address the problem of making a pre-trained reinforcement learning (RL) policy safety-aware by incorporating cost constraints without retraining it f
Pretty profound for an ipo filing: “For the entirety of its existence, human civilization has lived on a single celestial body: Earth. The current paradigm, in which human civilization is confined to
arXiv:2510.21583v2 Announce Type: replace Abstract: Recent Progress in post-training flow matching for text-to-image (T2I) generation with Group Relative Policy Optimization (GRPO) has demonstrated st
arXiv:2605.20355v1 Announce Type: cross Abstract: Skill atrophy, the gradual decline of human capability under AI assistance, poses a safety risk in shared-control of semi-autonomous systems, where op
arXiv:2605.20801v1 Announce Type: new Abstract: Adaptive robot navigation in dynamic environments requires policies that can reach the target reliably while producing efficient and stable trajectories
arXiv:2605.20482v1 Announce Type: new Abstract: Quadratic constraints (QCs) are widely used to characterize nonlinearities and uncertainties, but generic analytical characterizations can be conservati
arXiv:2605.20222v1 Announce Type: cross Abstract: Contextual combinatorial optimization (CCO) plays a critical role in decision-making under uncertainty, yet remains a significant challenge. We presen
arXiv:2605.20654v1 Announce Type: new Abstract: While Large Language Models (LLMs) demonstrate remarkable capabilities, they remain susceptible to sophisticated, multi-step jailbreak attacks that circ
arXiv:2605.20277v1 Announce Type: new Abstract: Medical vision-language models (VLMs) have rapidly advanced as general-purpose multimodal assistants, yet their deployment in 3D Computed Tomography (CT
arXiv:2605.21257v1 Announce Type: new Abstract: Planning through crowded environments under uncertain obstacle motions remains difficult, as stochastic interactions often induce overly conservative be
arXiv:2509.22963v3 Announce Type: replace Abstract: Reinforcement learning (RL) struggles to scale to large, combinatorial action spaces common in many real-world problems. This paper introduces a nov
arXiv:2602.07832v2 Announce Type: replace Abstract: Process rewards have been widely used in deep reinforcement learning to improve training efficiency, reduce variance, and prevent reward hacking. In
arXiv:2605.20708v1 Announce Type: new Abstract: Diffusion Transformers (DiTs) have become a de facto backbone of modern visual generation, and nearly every major axis of their design -- tokenization,
arXiv:2605.20721v1 Announce Type: new Abstract: Learning from implicit feedback in recommender systems is fundamentally challenged by pervasive label noise. While conventional denoising approaches oft
arXiv:2605.21186v1 Announce Type: new Abstract: Interpretability in object detection provides crucial confidence support for clinical auxiliary diagnosis. However, in tiny bacteria detection, traditio
arXiv:2605.20712v1 Announce Type: new Abstract: Automatic speech recognition replaces typing only when correction costs less than manual entry, a threshold determined by error types, not counts: fixin
arXiv:2601.00418v2 Announce Type: replace-cross Abstract: We propose the Consensus-Based Privacy-Preserving Data Distribution (CPPDD) framework, a lightweight and post-setup autonomous protocol for se
arXiv:2601.18577v2 Announce Type: replace Abstract: Modern video generators still struggle with complex physical dynamics, often falling short of physical realism. Existing approaches address this usi
arXiv:2605.20191v1 Announce Type: new Abstract: Modern Large Language Models (LLMs) have recently attracted much attention for their ability to simulate human behavior and generate text that reflects