Carlos's Debrief

May 19, 2026 04:00
37ArXiv Papers
45Web Findings
82Total Sources
Topics searched: Artificial Intelligence Machine Learning Large Language Models Security & Cybersecurity Cryptography Zero Knowledge Quantum Computing Crypto & Blockchain AI Agents & Reasoning AI Safety & Alignment
May 19, 2026 04:00 — Curated by Hermes Research Scout

⚡ Quick Summary

📰 News Headlines

📄 ArXiv Papers

🧠 LLMs & Agents 14

DashAttention: Differentiable and Adaptive Sparse Hierarchical Attention [1]
Yuxiang Huang, Nuno M. T. Gonçalves, Federico Alvetreti, Lei Li, Xu Han, Edoardo M. Ponti, André F. T. Martins, Marcos V. Treviso
Current hierarchical attention methods, such as NSA and InfLLMv2, select the top-k relevant key-value (KV) blocks based on coarse attention scores and subsequently apply fine-grained softmax attention on the selected tokens. However, the top-k operat…
cs.CL
Code as Agent Harness [2]
Xuying Ning, Katherine Tieu, Dongqi Fu, Tianxin Wei, Zihao Li, Yuanchen Bei, Jiaru Zou, Mengting Ai, Zhining Liu, Ting-Wei Li, Lingjie Chen, Yanjun Zh
Recent large language models (LLMs) have demonstrated strong capabilities in understanding and generating code, from competitive programming to repository-level software engineering. In emerging agentic systems, code is no longer only a target output…
cs.CL
ESI-Bench: Towards Embodied Spatial Intelligence that Closes the Perception-Action Loop [3]
Yining Hong, Jiageng Liu, Han Yin, Manling Li, Leonidas Guibas, Li Fei-Fei, Jiajun Wu, Yejin Choi
Spatial intelligence unfolds through a perception-action loop: agents act to acquire observations, and reason about how observations vary as a function of action. Rather than passively processing what is seen, they actively uncover what is unseen - o…
cs.CV
Actionable World Representation [4]
Kunqi Xu, Jitao Li, Jianglong Ye, Tianshu Tang, Isabella Liu, Sifei Liu, Xueyan Zou
Inspired by the emergent behaviors in large language models that generalized human intelligence, the research community is pursuing similar emergent capabilities within world models, with a emphasis on modeling the physical world. Within the scope of…
cs.AI
Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation [5]
Qianhao Yuan, Jie Lou, Xing Yu, Hongyu Lin, Le Sun, Xianpei Han, Yaojie Lu
Multimodal Large Language Models (MLLMs) still struggle with fine-grained visual understanding, where answers often depend on small but decisive evidence in the full image. We observe a regional-to-global perception gap: the same MLLM answers fine-gr…
cs.CV
What Does the AI Doctor Value? Auditing Pluralism in the Clinical Ethics of Language Models [6]
Payal Chandak, Victoria Alkin, David Wu, Maya Dagan, Taposh Dutta Roy, Maria Clara Saad Menezes, Ayush Noori, Nirali Somia, John S. Brownstein, Ran Ba
Medicine is inherently pluralistic. Principles such as autonomy, beneficence, nonmaleficence, and justice routinely conflict, and such ethical dilemmas often sharply divide reasonable physicians. Good clinical practice navigates these tensions in con…
cs.AI
Predictable Confabulations: Factual Recall by LLMs Scales with Model Size and Topic Frequency [7]
Matthew L. Smith, Jonathan P. Shock, Samuel T. Segun, Iyiola E. Olatunji, Tegawendé F. Bissyandé
While scaling laws govern aggregate large language model performance, no scaling law has linked factual recall to both model size and training-data composition. We evaluated 38 models on over 8,900 scholarly references evaluated by an automated refer…
cs.CL
DexHoldem: Playing Texas Hold'em with Dexterous Embodied System [8]
Feng Chen, Tianzhe Chu, Li Sun, Pei Zhou, Zhuxiu Xu, Shenghua Gao, Yuexiang Zhai, Yanchao Yang, Yi Ma
Evaluating embodied systems on real dexterous hardware requires more than isolated primitive skills: an agent must perceive a changing tabletop scene, choose a context-appropriate action, execute it with a dexterous hand, and leave the scene usable f…
cs.RO
Semantic Generative Tuning for Unified Multimodal Models [9]
Songsong Yu, Yuxin Chen, Ying Shan, Yanwei Li
Unified multimodal models (UMMs) strive to consolidate visual understanding and visual generation within a single architecture. However, prevailing training paradigms independently optimize understanding via sparse text signals and generation through…
cs.CV
Distilling Tabular Foundation Models for Structured Health Data [10]
Aditya Tanna, Nassim Bouarour, Mohamed Bouadi, Vinay Kumar Sankarapu, Pratinav Seth
Tabular foundation models (TFMs) achieve strong performance on health datasets, but their inference cost and infrastructure requirements limit practical use. We study whether their predictive behavior can be transferred to lightweight tabular models…
cs.LG
Can These Views Be One Scene? Evaluating Multiview 3D Consistency when 3D Foundation Models Hallucinate [17]
Soumava Paul, Prakhar Kaushik, Alan Yuille
Multiview 3D evaluation assumes that the images being scored are observations of one static 3D scene. This assumption can fail in NVS and sparse-view reconstruction: inputs or generated outputs may contain artifacts, outlier frames, repeated views, o…
cs.CV
Traditional statistical representations outperform generative AI in identifying expert peer reviewers [18]
Vicente Amado Olivo, Tereza Jerabkova, Jakub Klencki, John Carpenter, Mario Malički, Ferdinando Patat, Louis-Gregory Strolger, Wolfgang Kerzendorf
The exponential growth of scientific submissions has strained the peer review system. Despite the rapidly expanding global pool of researchers, this unprecedented scale has rendered the previous approach of manual expert identification unfeasible. Th…
cs.IR
WavFlow: Audio Generation in Waveform Space [19]
Feiyan Zhou, Luyuan Wang, Shoufa Chen, Zhe Wang, Zhiheng Liu, Yuren Cong, Xiaohui Zhang, Fanny Yang, Belinda Zeng
Modern audio generation predominantly relies on latent-space compression, introducing additional complexity and potential information loss. In this work, we challenge this paradigm with WavFlow, a framework that generates high-fidelity audio directly…
cs.SD
Aurora: Unified Video Editing with a Tool-Using Agent [20]
Yongsheng Yu, Ziyun Zeng, Zhiyuan Xiao, Zhenghong Zhou, Hang Hua, Wei Xiong, Jiebo Luo
Recent video editing models have converged on a unified conditioning design: a single diffusion transformer jointly consumes text, source video, and reference images, and one set of weights covers replacement, removal, style transfer, and reference-d…
cs.CV

⚙️ Machine Learning 6

A Readiness-Driven Runtime for Pipeline-Parallel Training under Runtime Variability [11]
Ruitao Liu, Xinyang Tian, Shuo Chen, Tingrui Zhang, Guang Yang, Alan Zhao, Wei Xu
Pipeline parallelism is a key technique for scaling large-model training, but modern workloads exhibit runtime variability in computation and communication. Existing pipeline systems typically consume static, profiled, or adaptively generated schedul…
cs.DC
SURGE: Approximation-free Training Free Particle Filter for Diffusion Surrogate [12]
Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu
Diffusion-based generative models increasingly rely on inference-time guidance, adding a drift term or reweighting mixture of experts, to improve sample quality on task-specific objectives. However, most existing techniques require repeated score or…
stat.ML
PIXLRelight: Controllable Relighting via Intrinsic Conditioning [13]
Miguel Farinha, Ronald Clark
We present PIXLRelight, a feed-forward approach for physically controllable single-image relighting. Existing methods either provide limited lighting control (e.g. through text or environment maps), accumulate errors when chaining inverse and forward…
cs.CV
General Preference Reinforcement Learning [14]
Muhammad Umer, Muhammad Ahmed Mohsin, Ahsan Bilal, Arslan Chaudhry, Andreas Haupt, Sanmi Koyejo, Emily Fox, John M. Cioffi
Post-training has split large language model (LLM) alignment into two largely disconnected tracks. Online reinforcement learning (RL) with verifiable rewards drives emergent reasoning on math and code but depends on a programmatic verifier that canno…
cs.LG
Learned Memory Attenuation in Sage-Husa Kalman Filters for Robust UAV State Estimation [15]
Kenan Majewski, Marcin Żugaj
Unmanned Aerial Vehicles in dynamic environments face telemetry outages, structural vibrations, and regime-dependent noise that invalidate the stationary covariance assumptions of classical Kalman filters. The Sage-Husa Kalman Filter (SHKF) estimates…
eess.SP
EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL [16]
Minrui Xu, Zilin Wang, Mengyi DENG, Zhiwei Li, Zhicheng Yang, Xiao Zhu, Yinhong Liu, Boyu Zhu, Baiyu Huang, Chao Chen, Heyuan Deng, Fei Mi, Lifeng Sha
Equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL) is bottlenecked by two challenges: the lack of scalable, robust execution environments and the scarcity of realistic training data that captures implicit human…
cs.CL

🔒 Security & Crypto 13

Sublinear Risk-Limiting Audits from Direct Ballot Selection and Statistical Ballot Manifests [21]
Benjamin Fuller, Abigail Harrison, Alexander Russell
Risk-limiting audits (RLAs) are post-election auditing procedures that rigorously guarantee a specified maximum probability that an incorrect electoral outcome will not be detected. Aside from ready access to physical ballots, known RLAs require a so…
cs.CR
A No-Defense Defense Against Gradient-Based Adversarial Attacks on ML-NIDS: Is Less More? [22]
Mohamed elShehaby, Ashraf Matrawy
Gradient-based adversarial attacks subtly manipulate inputs of Machine Learning (ML) models to induce incorrect predictions. This paper investigates whether careful architectural choices alone can yield an inherently robust Deep Neural Network (DNN)-…
cs.LG
Federated Naive Bayes with Real Mixture of Gaussians and Institutional Governance Regularization for Network Intrusion Detection [23]
Herrera Logroño, Edgar Oswaldo; López Rubio, Ezequiel, Ortiz de Lazcano Lobato, Juan Miguel
Federated learning for intrusion detection rests on a flawed premise: that every participating institution contributes equally to the shared model. In practice, a financial institution with mature security controls and low vulnerability exposure prod…
cs.CR
Learning to Look Benign: Targeted Evasion of Malware Detectors via API Import Injection [24]
Juozas Dautartas, Olga Kurasova, Juozapas Rokas Čypas, Viktor Medvedev
Machine learning-based malware detectors are widely deployed in antivirus and endpoint detection systems, yet their reliance on static features makes them vulnerable to adversarial manipulation. This paper investigates whether a malware sample can be…
cs.CR
Not What You Asked For: Typographic Attacks in Household Robot Manipulation [25]
Ali Iranmanesh, Peng Liu
Open-vocabulary embodied AI agents increasingly rely on vision-language models such as CLIP for object perception and task grounding. However, the shared embedding space that enables this flexibility introduces a structural vulnerability to typograph…
cs.CR
Overeager Coding Agents: Measuring Out-of-Scope Actions on Benign Tasks [26]
Yubin Qu, Ying Zhang, Yanjun Zhang, Gelei Deng, Yuekang Li, Leo Yu Zhang, Yi Liu
Coding agents now run autonomously with shell, file, and network privileges. When a user issues a benign request, the agent sometimes does more than asked: it deletes unrelated files, wipes a stale credentials backup, or rewrites configuration the us…
cs.SE
Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics [27]
Maciej Chrabąszcz, Aleksander Szymczyk, Marcin Sendera, Tomasz Trzciński, Sebastian Cygert
Large Reasoning Models (LRMs) introduce new opportunities for safety monitoring through their Chain of Thought (CoT) reasoning. However, CoT is not always faithful to the model's final output, undermining its reliability as a monitoring tool. To addr…
cs.CL
Bridging the Cybersecurity Gap Between Web2 and Web3 - An Incident-Based Analysis of Organizational and Application-Level Security Failures [28]
Tarkan Yavas, Arslan Brömme
The rapid adoption of Web3 infrastructures has led to a growing number of security incidents affecting cryptocurrency exchanges, custody services and blockchain-based platforms. While existing research predominantly focuses on vulnerabilities in smar…
cs.CR
Advancing Narrative Long Video Generation via Training-Free Identity-Aware Memory [32]
Jinzhuo Liu, Jiangning Zhang, Wencan Jiang, Yabiao Wang, Dingkang Liang, Zhucun Xue, Ran Yi, Yong Liu
Autoregressive video generation has improved rapidly in visual fidelity and interactivity, but it still suffers from long-term inconsistency and memory degradation. Most existing solutions either compress historical frames using predefined strategies…
cs.CV
Quantum Shannon theory made robust: a tale of three protocols for almost i.i.d. sources [33]
Filippo Girardi, Nilanjana Datta, Giacomo De Palma, Ludovico Lami
The asymptotic rates of information-theoretic protocols - including error exponents, compression rates, and channel capacities - are traditionally defined under the idealised assumption that the underlying resource (state or channel) is independent a…
quant-ph
Ranking Opinions with Few States in Population Protocols [34]
Tom-Lukas Breitkopf, Julien Dallot, Antoine El-Hayek, Stefan Schmid
Population protocols are a model of distributed computing where $n$ agents, each a simple finite-state machine, interact in pairs to solve a common task against a (adversarial) interaction scheduler. This model was intensively studied in recent years…
cs.DC
SkillGenBench: Benchmarking Skill Generation Pipelines for LLM Agents [35]
Yifan Zhou, Zhentao Zhang, Ziming Cheng, Shuo Zhang, Qizhen Lan, Zhangquan Chen, Zhi Yang, QianyuXu, Ronghao Chen, Huacan Wang, Sen Hu
As LLM agents are increasingly built around reusable skills, a central challenge is no longer only whether agents can use provided skills, but whether they can generate correct, reusable, and executable skills from repositories and documents. Existin…
cs.AI
Reversa: A Reverse Documentation Engineering Framework for Converting Legacy Software into Operational Specifications for AI Agents [36]
Sanderson Oliveira de Macedo, Ronaldo Martins da Costa
Legacy systems concentrate business rules, architectural decisions, and operational exceptions that often remain implicit in code, data, configuration, and maintenance practices. At the same time, language-model-based coding agents depend on reliab…
cs.SE

⚛️ Quantum 3

Robust Simulation Based Inference Through Robust Optimal Transport [29]
Peter Matthew Jacobs, Lekha Patel, Anirban Bhattacharya, Debdeep Pati
When a statistical model $\{P_θ : θ\in Θ\}$ lacks analytically tractable likelihoods, parametric statistical inference based on data generated from an unknown underlying distribution $P$ can still be performed as long as simulations from the model ar…
stat.ME
LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation [30]
Yukang Chen, Luozhou Wang, Wei Huang, Shuai Yang, Bohan Zhang, Yicheng Xiao, Ruihang Chu, Weian Mao, Qixin Hu, Shaoteng Liu, Yuyang Zhao, Huizi Mao, Y
We present LongLive-2.0, an NVFP4-based parallel infrastructure throughout the full training and inference workflow of long video generation, addressing speed and memory bottlenecks. For training, we introduce sequence-parallel autoregressive (AR) tr…
cs.CV
Spectral Progressive Diffusion for Efficient Image and Video Generation [31]
Howard Xiao, Brian Chao, Lior Yariv, Gordon Wetzstein
Diffusion models have been shown to implicitly generate visual content autoregressively in the frequency domain, where low-frequency components are generated earlier in the denoising process while high-frequency details emerge only in later timesteps…
cs.CV

🛡️ AI Safety 1

A universal framework to identify eccentric binary mergers: GW200105 case study [37]
Teagan A. Clarke, Isobel M. Romero-Shaw, Charlie Hoy, Jakob Stegmann, Paul D. Lasky, Eric Thrane
Orbital eccentricity in gravitational-wave signals from merging compact object binaries is a powerful indicator of their formation channel. Several binary black hole mergers and a neutron star--black hole merger have been reported to exhibit signs of…
astro-ph.HE

🌐 Web Findings

🦞 Lobste.rs 3

End of the semester [38]
Lobste.rs
A story on how I plan to spend my Holiday
Lobste.rs
The First CVE Wave: Signs That AI-Assisted Vulnerability Discovery Is Reshaping Disclosure Volumes [39]
Lobste.rs
Public CVE disclosure volumes are surging across major software suppliers and open source projects, and the evidence increasingly points to AI-assisted vulnerability discovery as t…
Lobste.rs
ChatGPT Won't Let You Type Until Cloudflare Reads Your React State. I Decrypted the Program That Does It [40]
Lobste.rs
Edit April 2, 2026: I've been getting inbound interest from researchers wanting to run their own queries. The MCP integration I use for my own research lets you analyze live mobile…
Lobste.rs

🤗 HuggingFace 32

FINESSE-Bench: A Hierarchical Benchmark Suite for Financial Domain Knowledge and Technical Analysis in Large Language Models [41]
HuggingFace
Large language models (LLMs) are increasingly being applied to financial analysis, reporting, investment decision support, risk management, compliance, and professional training. H…
HuggingFace
SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training [42]
HuggingFace
Diffusion models have been widely studied for removing unsafe content learned during pre-training. Existing methods require expensive supervised data, either unsafe-text paired wit…
HuggingFace
A2RBench: An Automatic Paradigm for Formally Verifiable Abstract Reasoning Benchmark Generation [43]
HuggingFace
Abstract reasoning ability reflects the intelligence and generalization capacity of LLMs to extract and apply abstract rules. However, accurately measuring this ability remains cha…
HuggingFace
GRASP: Learning to Ground Social Reasoning in Multi-Person Non-Verbal Interactions [44]
HuggingFace
Understanding social interactions requires reasoning over subtle non-verbal cues, yet current multimodal large language models (MLLMs) often fail to identify who interacts with who…
HuggingFace
Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics [45]
HuggingFace
Large Reasoning Models (LRMs) introduce new opportunities for safety monitoring through their Chain of Thought (CoT) reasoning. However, CoT is not always faithful to the model's f…
HuggingFace
SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolution [46]
HuggingFace
Long-horizon LLM agents leave traces that could become reusable experience, but raw trajectories are noisy and hard to govern. We treat Agent Skills as an experience schema that co…
HuggingFace
MixSD: Mixed Contextual Self-Distillation for Knowledge Injection [47]
HuggingFace
Supervised fine-tuning (SFT) is widely used to inject new knowledge into language models, but it often degrades pretrained capabilities such as reasoning and general-domain perform…
HuggingFace
Code-as-Room: Generating 3D Rooms from Top-Down View Images via Agentic Code Synthesis [48]
HuggingFace
Designing realistic and functional 3D indoor rooms is essential for a wide range of applications, including interior design, virtual reality, gaming, and embodied AI. While recent…
HuggingFace
From Runnable to Shippable: Multi-Agent Test-Driven Development for Generating Full-Stack Web Applications from Requirements [49]
HuggingFace
Coding agents can generate web applications from natural-language descriptions, yet a recent benchmark study shows that generated applications fail to meet functional requirements…
HuggingFace
StableVLA: Towards Robust Vision-Language-Action Models without Extra Data [50]
HuggingFace
It is infeasible to encompass all possible disturbances within the training dataset. This raises a critical question regarding the robustness of Vision-Language-Action (VLA) models…
HuggingFace
Incantation: Natural Language as the Action Interface for Multi-Entity Video World Models [51]
HuggingFace
Modern interactive video world models have achieved impressive visual fidelity, yet lack fine-grained multi-entity control and cross-entity, cross-world generalization. We trace th…
HuggingFace
AgentKernelArena: Generalization-Aware Benchmarking of GPU Kernel Optimization Agents [52]
HuggingFace
GPU kernel optimization is increasingly critical for efficient deep learning systems, but writing high-performance kernels still requires substantial low-level expertise. Recent AI…
HuggingFace
AtlasVA: Self-Evolving Visual Skill Memory for Teacher-Free VLM Agents [53]
HuggingFace
Vision-language model (VLM) agents increasingly rely on memory-augmented reinforcement learning to reuse experience across long-horizon tasks, yet most existing frameworks store me…
HuggingFace
Actionable World Representation [54]
HuggingFace
Inspired by the emergent behaviors in large language models that generalized human intelligence, the research community is pursuing similar emergent capabilities within world model…
HuggingFace
Code as Agent Harness [55]
HuggingFace
Recent large language models (LLMs) have demonstrated strong capabilities in understanding and generating code, from competitive programming to repository-level software engineerin…
HuggingFace
CompactAttention: Accelerating Chunked Prefill with Block-Union KV Selection [56]
HuggingFace
Chunked prefill has become a widely adopted serving strategy for long-context large language models, but efficient attention computation in this regime remains challenging. Existin…
HuggingFace
OProver: A Unified Framework for Agentic Formal Theorem Proving [57]
HuggingFace
Recent progress in formal theorem proving has benefited from large-scale proof generation and verifier-aware training, but agentic proving is rarely integrated into prover training…
HuggingFace
Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces [58]
HuggingFace
The deployment of Large Language Models (LLMs) as autonomous economic agents introduces systemic risks that extend beyond individual capability failures. As agents transition to di…
HuggingFace
Lance: Unified Multimodal Modeling by Multi-Task Synergy [59]
HuggingFace
We present Lance, a lightweight native unified model supporting multimodal understanding, generation, and editing for both images and videos. Rather than relying on model capacity…
HuggingFace
Post-Trained MoE Can Skip Half Experts via Self-Distillation [60]
HuggingFace
Mixture-of-Experts (MoE) scales language models efficiently through sparse expert activation, and its dynamic variant further reduces computation by adjusting the activated experts…
HuggingFace
Targeted Neuron Modulation via Contrastive Pair Search [61]
HuggingFace
Language models are instruction-tuned to refuse harmful requests, but the mechanisms underlying this behavior remain poorly understood. Popular steering methods operate on the resi…
HuggingFace
LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation [62]
HuggingFace
We present LongLive-2.0, an NVFP4-based parallel infrastructure throughout the full training and inference workflow of long video generation, addressing speed and memory bottleneck…
HuggingFace
AI for Auto-Research: Roadmap & User Guide [63]
HuggingFace
AI-assisted research is crossing a threshold: fully automated systems can now generate research papers for as little as $15, while long-horizon agents can execute experiments, draf…
HuggingFace
KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration [64]
HuggingFace
Aligning streaming autoregressive (AR) video generators with human preferences is challenging. Existing reinforcement learning methods predominantly rely on noise-based exploration…
HuggingFace
Where Should Diffusion Enter a Language Model? Geometry-Guided Hidden-State Replacement [65]
HuggingFace
Continuous diffusion language models lag behind autoregressive transformers, partly because diffusion is applied in spaces poorly suited to language denoising and token recovery. W…
HuggingFace
Measuring Maximum Activations in Open Large Language Models [66]
HuggingFace
The dynamic range of activations is a first-order constraint for low-bit quantization, activation scaling, and stable LLM inference. Prior work characterized outlier features and m…
HuggingFace
EndPrompt: Efficient Long-Context Extension via Terminal Anchoring [67]
HuggingFace
Extending the context window of large language models typically requires training on sequences at the target length, incurring quadratic memory and computational costs that make lo…
HuggingFace
Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Models [68]
HuggingFace
Large Reasoning Models (LRMs) achieve strong performance by generating long chains of thought (CoT), but often overthink, continuing to reason after a solution has already stabiliz…
HuggingFace
LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs [69]
HuggingFace
The fundamental challenge in scaling Video Large Language Models (Video LLMs) to long-form video lies in managing the explosion of visual-token context length. Existing strategies…
HuggingFace
Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use [70]
HuggingFace
Large language models (LLMs) increasingly act as autonomous agents that must decide when to answer directly vs. when to invoke external tools. Prior work studying adaptive tool use…
HuggingFace
SNLP: Layer-Parallel Inference via Structured Newton Corrections [71]
HuggingFace
Autoregressive language models execute Transformer layers sequentially, creating a latency bottleneck that is not removed by conventional tensor or pipeline parallelism. We study w…
HuggingFace
Geometric Phase Transition Enables Extreme Hippocampal Memory Capacity [72]
HuggingFace
Memory systems can store vastly different amounts of information despite similar hardware constraints. Here, we show that superior spatial memory emerges from a discrete stiffening…
HuggingFace

🎓 Google Scholar 10

Human-computer interaction (HCI) technologies in socially-enabled artificial intelligence [73]
Google Scholar
Human-computer interaction (HCI) technologies in socially-enabled artificial intelligence
Google Scholar
Early identification of breakthrough technologies: Insights from science-driven innovations [74]
Google Scholar
Early identification of breakthrough technologies: Insights from science-driven innovations
Google Scholar
Catalyst breakthroughs in methane dry reforming: Employing machine learning for future advancements [75]
Google Scholar
Catalyst breakthroughs in methane dry reforming: Employing machine learning for future advancements
Google Scholar
Large language models (LLM) in computational social science: prospects, current state, and challenges [76]
Google Scholar
Large language models (LLM) in computational social science: prospects, current state, and challenge
Google Scholar
A comprehensive review of generative AI techniques and their impact on cybersecurity [77]
Google Scholar
In today’s digital landscape, cybersecurity stands as the cornerstone of defense against an ever-evolving array of threats, ranging from ransomware attacks to state-sponsored espio…
Google Scholar
Securing the future: exploring post-quantum cryptography for authentication and user privacy in IoT devices [78]
Google Scholar
Cluster Computing - With the emergence of quantum computers, traditional cryptographic methods are vulnerable to attacks, emphasizing the need for post-quantum cryptography to secu…
Google Scholar
Quantum machine learning: A comprehensive review of integrating AI with quantum computing for computational advancements [79]
Google Scholar
Quantum machine learning: A comprehensive review of integrating AI with quantum computing for comput
Google Scholar
A systematic literature review on blockchain-based smart contracts: Platforms, applications, and challenges [80]
Google Scholar
A systematic literature review on blockchain-based smart contracts: Platforms, applications, and cha
Google Scholar
Smart Contracts, Blockchain, and Health Policies: Past, Present, and Future [81]
Google Scholar
Smart Contracts, Blockchain, and Health Policies: Past, Present, and Future
Google Scholar
Controllable safety alignment: Inference-time adaptation to diverse safety requirements [82]
Google Scholar
Jingyu Zhang, Ahmed Elgohary Ghoneim, Ahmed Magooda, Daniel Khashabi, Ben Van Durme
Google Scholar

🔗 All Sources

  1. [1] DashAttention: Differentiable and Adaptive Sparse Hierarchical Attenti
  2. [2] Code as Agent Harness
  3. [3] ESI-Bench: Towards Embodied Spatial Intelligence that Closes the Perce
  4. [4] Actionable World Representation
  5. [5] Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Po
  6. [6] What Does the AI Doctor Value? Auditing Pluralism in the Clinical Ethi
  7. [7] Predictable Confabulations: Factual Recall by LLMs Scales with Model S
  8. [8] DexHoldem: Playing Texas Hold'em with Dexterous Embodied System
  9. [9] Semantic Generative Tuning for Unified Multimodal Models
  10. [10] Distilling Tabular Foundation Models for Structured Health Data
  11. [11] A Readiness-Driven Runtime for Pipeline-Parallel Training under Runtim
  12. [12] SURGE: Approximation-free Training Free Particle Filter for Diffusion
  13. [13] PIXLRelight: Controllable Relighting via Intrinsic Conditioning
  14. [14] General Preference Reinforcement Learning
  15. [15] Learned Memory Attenuation in Sage-Husa Kalman Filters for Robust UAV
  16. [16] EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthe
  17. [17] Can These Views Be One Scene? Evaluating Multiview 3D Consistency when
  18. [18] Traditional statistical representations outperform generative AI in id
  19. [19] WavFlow: Audio Generation in Waveform Space
  20. [20] Aurora: Unified Video Editing with a Tool-Using Agent
  21. [21] Sublinear Risk-Limiting Audits from Direct Ballot Selection and Statis
  22. [22] A No-Defense Defense Against Gradient-Based Adversarial Attacks on ML-
  23. [23] Federated Naive Bayes with Real Mixture of Gaussians and Institutional
  24. [24] Learning to Look Benign: Targeted Evasion of Malware Detectors via API
  25. [25] Not What You Asked For: Typographic Attacks in Household Robot Manipul
  26. [26] Overeager Coding Agents: Measuring Out-of-Scope Actions on Benign Task
  27. [27] Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning
  28. [28] Bridging the Cybersecurity Gap Between Web2 and Web3 - An Incident-Bas
  29. [29] Robust Simulation Based Inference Through Robust Optimal Transport
  30. [30] LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generati
  31. [31] Spectral Progressive Diffusion for Efficient Image and Video Generatio
  32. [32] Advancing Narrative Long Video Generation via Training-Free Identity-A
  33. [33] Quantum Shannon theory made robust: a tale of three protocols for almo
  34. [34] Ranking Opinions with Few States in Population Protocols
  35. [35] SkillGenBench: Benchmarking Skill Generation Pipelines for LLM Agents
  36. [36] Reversa: A Reverse Documentation Engineering Framework for Converting
  37. [37] A universal framework to identify eccentric binary mergers: GW200105 c
  38. [38] End of the semester
  39. [39] The First CVE Wave: Signs That AI-Assisted Vulnerability Discovery Is
  40. [40] ChatGPT Won't Let You Type Until Cloudflare Reads Your React State. I
  41. [41] FINESSE-Bench: A Hierarchical Benchmark Suite for Financial Domain Kno
  1. [42] SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Train
  2. [43] A2RBench: An Automatic Paradigm for Formally Verifiable Abstract Reaso
  3. [44] GRASP: Learning to Ground Social Reasoning in Multi-Person Non-Verbal
  4. [45] Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning
  5. [46] SkillsVote: Lifecycle Governance of Agent Skills from Collection, Reco
  6. [47] MixSD: Mixed Contextual Self-Distillation for Knowledge Injection
  7. [48] Code-as-Room: Generating 3D Rooms from Top-Down View Images via Agenti
  8. [49] From Runnable to Shippable: Multi-Agent Test-Driven Development for Ge
  9. [50] StableVLA: Towards Robust Vision-Language-Action Models without Extra
  10. [51] Incantation: Natural Language as the Action Interface for Multi-Entity
  11. [52] AgentKernelArena: Generalization-Aware Benchmarking of GPU Kernel Opti
  12. [53] AtlasVA: Self-Evolving Visual Skill Memory for Teacher-Free VLM Agents
  13. [54] Actionable World Representation
  14. [55] Code as Agent Harness
  15. [56] CompactAttention: Accelerating Chunked Prefill with Block-Union KV Sel
  16. [57] OProver: A Unified Framework for Agentic Formal Theorem Proving
  17. [58] Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces
  18. [59] Lance: Unified Multimodal Modeling by Multi-Task Synergy
  19. [60] Post-Trained MoE Can Skip Half Experts via Self-Distillation
  20. [61] Targeted Neuron Modulation via Contrastive Pair Search
  21. [62] LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generati
  22. [63] AI for Auto-Research: Roadmap & User Guide
  23. [64] KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semant
  24. [65] Where Should Diffusion Enter a Language Model? Geometry-Guided Hidden-
  25. [66] Measuring Maximum Activations in Open Large Language Models
  26. [67] EndPrompt: Efficient Long-Context Extension via Terminal Anchoring
  27. [68] Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reas
  28. [69] LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLM
  29. [70] Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Too
  30. [71] SNLP: Layer-Parallel Inference via Structured Newton Corrections
  31. [72] Geometric Phase Transition Enables Extreme Hippocampal Memory Capacity
  32. [73] Human-computer interaction (HCI) technologies in socially-enabled arti
  33. [74] Early identification of breakthrough technologies: Insights from scien
  34. [75] Catalyst breakthroughs in methane dry reforming: Employing machine lea
  35. [76] Large language models (LLM) in computational social science: prospects
  36. [77] A comprehensive review of generative AI techniques and their impact on
  37. [78] Securing the future: exploring post-quantum cryptography for authentic
  38. [79] Quantum machine learning: A comprehensive review of integrating AI wit
  39. [80] A systematic literature review on blockchain-based smart contracts: Pl
  40. [81] Smart Contracts, Blockchain, and Health Policies: Past, Present, and F
  41. [82] Controllable safety alignment: Inference-time adaptation to diverse sa
Generated by Hermes Agent • arXiv • Google Scholar • Lobste.rs • Hacker News • HuggingFace