Tail-Likelihood Reinforcement Learning
arXiv:2609.02987v1 Announce Type: new Abstract: Reinforcement learning typically optimizes average reward. For generative policies, the average can hide an important distinction: two policies can achi
Knowledge catalogue
arXiv:2609.02987v1 Announce Type: new Abstract: Reinforcement learning typically optimizes average reward. For generative policies, the average can hide an important distinction: two policies can achi
arXiv:2609.04071v1 Announce Type: cross Abstract: Pathology foundation models improve transferable representation learning for histopathology, but recent gains often rely on encoders with hundreds of
arXiv:2602.06638v3 Announce Type: replace-cross Abstract: Predictive confidence serves as a foundational control signal in mission-critical systems, directly governing risk-aware logic such as escalat
arXiv:2609.04203v1 Announce Type: new Abstract: We introduce S^3T (Self-Supervised Self-Distillation over Time), which, to the best of our knowledge, is the first fully self-contained framework for co
arXiv:2609.03302v1 Announce Type: new Abstract: We present a unified computational approach to tensor-based morphometry in detecting the brain surface shape differences between two clinical groups bas
arXiv:2609.04148v1 Announce Type: new Abstract: As terminal-based code agents become prevalent, agent trajectories have accumulated at scale, while realistic, executable environments remain scarce. Ho
arXiv:2609.03622v1 Announce Type: cross Abstract: Test-time adaptation (TTA) offers a promising direction for improving speech enhancement models under mismatched acoustic conditions, without requirin
arXiv:2609.03585v1 Announce Type: new Abstract: Thermal infrared imaging offers reliable perception in darkness and adverse weather, but thermal datasets remain scarce, motivating extensive work on tr
arXiv:2609.03231v1 Announce Type: new Abstract: The ambition of the 2025 PNPL competition (Landau et al., 2025) was to launch a multi-year curriculum for non-invasive speech decoding. Designed to prog
The AI analysis agents that helped investigate the rogue OpenAI agents kept taking their side. That detail was buried in @dylfreed's New York Times coverage of the recent @METR_Evals investigation of
arXiv:2609.03218v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly use user context such as memory, profiles, and role prompts to personalize their responses. This personalizati
arXiv:2609.03586v1 Announce Type: new Abstract: Audio-video diffusion models rely on cross-modal attention to coordinate text, sound, and visual content, yet this same mechanism can introduce subtle a
arXiv:2609.04009v1 Announce Type: cross Abstract: Noisy annotations pose a significant challenge for supervised deep learning, as neural networks rely on large-scale, high-quality labeled data whose c
arXiv:2609.03425v1 Announce Type: cross Abstract: Humans are the transport layer between AI systems, losing context at every hop. We present the Civilization Framework, whose addressable party is the
arXiv:2609.04047v1 Announce Type: cross Abstract: Background: Researchers increasingly use repeated identical prompts to audit stochastic variation in large language model (LLM) brand recommendations,
arXiv:2609.04005v1 Announce Type: new Abstract: We present an alternative characterization of the occupancy measure of reinforcement learning, obtained by embedding the planning criterion into the dyn
arXiv:2609.02959v1 Announce Type: cross Abstract: What does a language model predict when it has few clues? The answer lurks in its unembedding geometry: a single direction of the unembedding matrix e
arXiv:2609.03090v1 Announce Type: new Abstract: Continuous chain-of-thought models compress reasoning into latent tokens. Matrix-valued variants, which route each latent token through a d x d matrix b
arXiv:2609.04046v1 Announce Type: cross Abstract: What can a single layer of self-attention compute? We study head complexity: the minimum number of attention heads required to compute a function in a
arXiv:2609.03829v1 Announce Type: cross Abstract: Few-shot fine-grained image classification (FSFGIC) aims to classify similar images with limited labeled examples. This work highlights the critical y
arXiv:2609.03652v1 Announce Type: new Abstract: Synthetic data augmentation has become a common strategy for addressing class imbalance in NLP, but most approaches focus on the quantity and diversity
arXiv:2609.04135v1 Announce Type: new Abstract: AI agents are increasingly being developed and deployed across organizations using heterogeneous agent-development frameworks, AI models, tool interface
I got access to GPT-6 Astra this afternoon, so naturally I used it to generate SVGs of pelicans riding bicycles - at low, medium, high, xhigh and max reasoning levels (Astra doesn't support reasoning=
The preview releases of Solaris and GWM 2 Worlds this week are just a glimpse of what happens when video world models are scaled significantly and paired with real-time inference. It's a real-time sim
arXiv:2609.03456v1 Announce Type: cross Abstract: Artificial intelligence (AI) is increasingly used to augment software engineering (SE) workflows. While code generation remains the main use case, org
the reception is unlike anything i thought possible for a 2026 OAI launch https://x.com/latentspacepod/status/2095757111854866639 for the first time in a year, @OpenAI's model launch has been better r
arXiv:2609.04110v1 Announce Type: new Abstract: Seeing frames in order does not mean representing time. Modern VideoLMs receive ordered video streams, yet their main supervision acts on generated text
the united states split the atom, landed on the moon, beat polio, doubled the global wheat yield, sequenced the genome, and put a computer in every pocket. we've taken steps backwards too, like doom s
arXiv:2609.03633v1 Announce Type: cross Abstract: Chain-of-thought (CoT) reasoning improves large reasoning models (LRMs) on complex tasks but often produces long, redundant traces. Recent training-fr
This has steadily expanded in scope over the last few hours, with multiple independent investigators finding that dozens of wikis may have been edited by this same swarm since May. The most interestin
⚠️ This is really scary. Proxies for OpenAI are laying the groundwork for saying the rest of us should suck up rogue AI risk and quit our bitching. Because they know they are on the verge of losing co
This work introduces diffusion-augmented LLMs, a new class of models. They first suggest that speculative decoding needs a separate draft model. Diffusion LLMs give up the quality of the model they re
arXiv:2606.18338v2 Announce Type: replace Abstract: The search for life beyond Earth will depend on detecting faint signatures in the atmospheres of potentially habitable exoplanets. Interpreting thos
arXiv:2609.03383v1 Announce Type: new Abstract: Graph-based policy optimization improves credit assignment for long-horizon LLM agents by organizing rollout trajectories into state-transition graphs.
arXiv:2609.03358v1 Announce Type: new Abstract: Numerical simulation of dynamical systems is usually organized as a causal march through time: each state is computed from the previous one. We explore
arXiv:2609.03410v1 Announce Type: new Abstract: Idiomatic expressions are an integral part of natural language, reflecting cultural nuances and posing unique challenges for computational models, parti
arXiv:2609.04202v1 Announce Type: new Abstract: While data-driven 3D shape correspondence estimation has recently seen substantial progress, robust matching under partial observations and strong non-i
arXiv:2609.03620v1 Announce Type: cross Abstract: Audio deepfake detection is commonly formulated as clip-level binary classification of single-domain audio. However, real-world manipulated audio can
arXiv:2609.03688v1 Announce Type: new Abstract: Pairwise preference labels rank complete images, yet Diffusion-DPO applies their effect over many spatial and denoising-time coordinates. For attention-
Being a person obsessed with testing new models that come out, times are really insane for me. Tested different kinds of TTS and voice cloning models but none of them gets it right in terms of emotion
arXiv:2609.03704v1 Announce Type: new Abstract: This paper presents an integrated approach to modeling human competencies by combining the theoretical foundations of cognitive architectures with princ
arXiv:2609.02990v1 Announce Type: cross Abstract: To scale up collective decision-making, participatory democracy platforms such as Polis and Remesh enable online deliberation among thousands of parti
arXiv:2609.03565v1 Announce Type: cross Abstract: Action-conditioned JEPA world models enable planning toward visually specified goals without reconstructing future pixels, yet latent prediction alone
arXiv:2609.03927v1 Announce Type: new Abstract: For robots to operate reliably in real-world environments, they need to perceive their surroundings, act, and reason about the consequences of those act
arXiv:2609.03501v1 Announce Type: cross Abstract: Mixture-of-experts (MoE) architectures increase model capacity by combining a collection of expert predictors through input-dependent routing, while o
arXiv:2601.04736v2 Announce Type: replace Abstract: Despite remarkable capability in multi-modal understanding, deploying Multi-modal Large Language Models (MLLMs) in open-ended conversational scenari
arXiv:2609.03938v1 Announce Type: new Abstract: While HTN planning has received significant attention in recent years, support for numerical reasoning remains very limited. In this paper, we investiga
arXiv:2609.02928v1 Announce Type: cross Abstract: Modern multi-agent systems are increasingly deployed at scale over large populations of agents in settings such as ad-auctions, traffic routing, and r
arXiv:2609.02991v1 Announce Type: new Abstract: Learned graph simulators provide an efficient alternative to high-fidelity solvers for granular dynamics. However, granular motion depends strongly on i
arXiv:2507.03887v1 Announce Type: cross Abstract: Recent advances in Text-To-Speech (TTS) technology have enabled synthetic speech to mimic human voices with remarkable realism, raising significant se
arXiv:2609.03561v1 Announce Type: new Abstract: Existing radar-LiDAR fusion methods rely on fixed residual weights, even though the informativeness of radar Doppler and LiDAR geometry is scan- and dir
arXiv:2609.03797v1 Announce Type: new Abstract: Long-term human-AI interaction is difficult because the information that guides inference is updated implicitly by the model and is not directly inspect
arXiv:2609.04048v1 Announce Type: cross Abstract: Neural machine translation (NMT) systems typically produce a single output per input, obscuring the alternative decision trajectories implicitly avail
arXiv:2609.03427v1 Announce Type: cross Abstract: Path representation learning (PRL) for road networks has received increasing research attention, due to various path-related applications. Existing wo
arXiv:2609.03480v1 Announce Type: cross Abstract: We map tree species across Denmark using National Forest Inventory plots and EO data, while evaluating the potential of foundation models for large-sc
arXiv:2609.03641v1 Announce Type: new Abstract: Vector-quantization based image compression has achieved strong rate--distortion performance, yet most of them still produce a separate compressed repre
arXiv:2609.03534v1 Announce Type: cross Abstract: 3D Gaussian Splatting has become a de facto scene representation for novel view synthesis, yet robustly learning 3D Gaussian primitives from visual in
Tulip uses two techniques to lower latency and boost throughput. CUDA graphs pre‑record GPU work so it launches in one call, reducing CPU overhead. LazyTensors tracks results asynchronously, letting t
arXiv:2609.03955v1 Announce Type: new Abstract: Reinforcement learning (RL) has substantially advanced code generation with large language models (LLMs) through executable feedback. The feedback for c
arXiv:2609.03775v1 Announce Type: new Abstract: Typological features are widely used in multilingual NLP, and the prediction of such features holds downstream utility. However, existing methods to pre