Carlos's Debrief

⚡ Quick Summary

Your Privacy My Cloak: Backdoor Attacks on Differentially Private Federated Learning: Prior research suggests that differential privacy (DP) inherently enhances the robustness of federated learning (FL) against backdoor attacks. In this paper, we challenge this assumption.
Selection Without Signal, Recovery Through Expression: A Measurement Study of Post-Hoc Falsification Operators for Frozen Small Code Models: Frozen small code models (<=1.5B parameters, run locally without fine-tuning) suit offline and privacy-constrained use, but often emit plausible-but-wrong programs. A natural remedy is a post-hoc operator that selects, verifies, repairs,…
Stable Menus of Public Goods: AI-Enabled Progress: Using an open problem from the EC 2025 paper "Stable Menus of Public Goods" as a testbed, we conduct experiments to understand the effectiveness of different AI-for-EconCS research workflows. Specifically, we study three questions: Does…
Automated jailbreak attack targeting multiple defense strategies: Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks. However, their safety remains a critical concern due to their susceptibility to adversarial prompt-based attacks.
Cross-Silo De-Anonymization Under Local Differential Privacy: Threat Model, Phase Transition, and Coordination Necessity: When a person's records appear in k independent data silos, each protected by (epsilon, delta)-differential privacy, standard composition yields a valid (k*epsilon, k*delta)-DP guarantee for the joint output. This worst-case bound,…

📰 News Headlines

JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence: Many moments in the real world do not wait for a user to ask.
DreamX-World 1.0: A General-Purpose Interactive World Model: DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation.
FastContext: Training Efficient Repository Explorer for Coding Agents: Large Language Model (LLM) coding agents have achieved strong results on software engineering tasks, yet repository exploration remains a major bottleneck: locating relevant code consumes substantial…
Who Should Lead Decoding Now? Tracking Reliable Trajectories for Ensembling Masked Diffusion Language Models: Masked Diffusion Language Models (MDLMs) have emerged as a distinct paradigm for sequence generation.
VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models: This technical report introduces VibeThinker-3B, a compact dense model with 3B parameters developed to investigate how far verifiable reasoning can be pushed within a strictly small-model regime.

🗡️ Relevant to your Katana work 2

Automated jailbreak attack targeting multiple defense strategies [28]
arXiv
BadWorld: Adversarial Attacks on World Models [71]
huggingface_papers · ▲ 13

📄 ArXiv Papers

🧠 Manipulation / Human / Geometric 6

Geometric Action Model for Robot Policy Learning [12]

Jisang Han, Seonghu Jeon, Jaewoo Jung, René Zurbrügg, Honggyu An, Tifanny Portela, Marco Hutter, Marc Pollefeys, Seungryong Kim, Sunghwan Hong

Generalist robot policies must follow user instructions while reasoning about how objects, cameras, and robot actions interact in the 3D physical world. Recent vision-language-action models (VLAs)…

cs.ROcs.CVcs.LG

ROVE: Unlocking Human Interventions for Humanoid Manipulation via Reinforcement Learning [44]

Wei Xiao, Weiliang Tang, Yuying Ge, Hui Zhou, Yao Mu, Li Zhang, Yixiao Ge

Human interventions provide crucial corrective signals for post-training Vision-Language-Action (VLA) models. However, enabling seamless humanoid interventions is a formidable systems challenge due…

cs.ROcs.LG

Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation [36]

Jie Zhang, Xiaoyue Chen, Anzhe Chen, Chenxu Lv, Deqing Li, Gengze Zhou, Hang Yin, Haoqi Yuan, Haoyang Li, Jiahao Li, Jiazhao Zhang, Jingren Zhou, Kaiyuan Gao, Kun Yan, Lihan Jiang, Ningyuan Tang, Pei Lin, Qihang Peng, Shengming Yin, Tianhe Wu, Tianyi Yan, Xiao Xu, Yan Shu, Yanran Zhang, Ye Wang, Yi Wang, Yilei Chen, Yixian Xu, Yiyang Huang, Yuxiang Chen, Zekai Zhang, Zhendong Wang, Zhixing Lei, Zhixuan Liang, Zihao Liu, Zikai Zhou, Xiong-Hui Chen, Chenfei Wu

We introduce Qwen-RobotWorld, a language-conditioned video world model for embodied intelligence. With natural language as a unified action interface, it predicts physically grounded future visual…

cs.CV

T-Rex: Tactile-Reactive Dexterous Manipulation [19]

Dantong Niu, Zhuoyang Liu, Zekai Wang, Boning Shao, Zhao-Heng Yin, Anirudh Pai, Yuvan Sharma, Stefano Saravalle, Ruijie Zheng, Jing Wang, Ryan Punamiya, Mengda Xu, Yuqi Xie, Yunfan Jiang, Letian Fu, Konstantinos Kallidromitis, Matteo Gioia, Junyi Zhang, Jiaxin Ge, Haiwen Feng, Fabio Galasso, Wei Zhan, David M. Chan, Yutong Bai, Roei Herzig, Jiahui Lei, Fei-Fei Li, Ken Goldberg, Jitendra Malik, Pieter Abbeel, Yuke Zhu, Danfei Xu, Jim, Fan, Trevor Darrell

The ability to react dynamically to tactile signals has long been considered crucial to agile human-level dexterity. Yet contemporary learning-based Vision-Language-Action (VLA) models for robotic…

cs.RO

Human Universal Grasping [20]

Kevin Yuanbo Wu, Tianxing Zhou, Isaac Tu, Billy Yan, Irmak Guzey, David Fouhey, Dandan Shan, Lerrel Pinto

Humans can grasp objects effortlessly, whereas multi-fingered robots are far from this level of generality. We argue that the most natural source of robot grasping data is from humans, who pick up…

cs.RO

R2RDreamer: 3D-aware Data Augmentation for Spatially-generalized 2D Manipulation Policies [49]

Xiuwei Xu, Haowen Sun, Angyuan Ma, Yiwei Zhang, Zhenyu Wu, Xiaofeng Wang, Bingyao Yu, Zheng Zhu, Jie Zhou, Jiwen Lu

Spatial generalization is critical for imitation-learned manipulation policies, but achieving it typically requires scaling demonstrations across diverse object poses, robot configurations, and…

cs.ROcs.CV

🧠 Reinforcement / When / Doubt 5

ExpRL: Exploratory RL for LLM Mid-Training [16]

Violet Xiang, Amrith Setlur, Chase Blagden, Nick Haber, Aviral Kumar

Sparse reward reinforcement learning (RL) has become a standard tool for improving LLM reasoning, but its success depends critically on the coverage present in the base model. In practice, models are…

cs.LG

The Value Axis: Language Models Encode Whether They're on the Right Track [18]

Nick Jiang, Isaac Kauvar, Jack Lindsey

We investigate whether language models internally track the value of their current trajectory, defined as the likelihood that their ongoing strategy will achieve their goals. Using synthetic,…

cs.CL

DEEPRUBRIC: Evidence-Tree Rubric Supervision for Efficient Reinforcement Learning of Deep Research Agents [50]

Minghang Zhu, Chuyang Wei, Junhao Xu, Yilin Cheng, Zhumin Chen, Jiyan He

Deep research agents synthesize long-form reports by searching and reasoning over retrieved evidence. Reinforcement learning with rubric-based rewards improves these agents by optimizing them against…

cs.CL

Context-Aware RL for Agentic and Multimodal LLMs [21]

Peiyang Xu, Bangzheng Li, Sijia Liu, Karthik R. Narasimhan, Pramod Viswanath, Prateek Mittal, Xingyu Fu

Large language models (LLMs) often fail when answering requires identifying a small but decisive piece of evidence within a long or complex context, such as a single line in a tool trace or a subtle…

cs.CLcs.CV

When in Doubt, Plan It Out: Committed Small Language Model Deliberation for Reactive Reinforcement Learning [8]

Nathan Gavenski, Juarez Monteiro, Francisco Galuppo, Adriano Veloso, Odinaldo Rodrigues

Reinforcement Learning (RL) policies often degrade in unfamiliar environments because they lack explicit deliberation. We propose Plan, Align, Commit, Think (PACT), a hybrid architecture that…

cs.AIcs.LG

🧠 Context / Tokenpilot / Cache-efficient 2

TokenPilot: Cache-Efficient Context Management for LLM Agents [4]

Buqiang Xu, Zirui Xue, Dianmou Chen, Chenyang Fu, Chiyu Wu, Caiying Huang, Chen Jiang, Jizhan Fang, Xinle Deng, Yijun Chen, Yunzhi Yao, Xuehai Wang, Jin Shang, Gong Yu, Ningyu Zhang

As LLM agents are deployed in long-horizon sessions, context accumulation drives up inference costs. Existing approaches utilize text pruning or dynamic memory eviction to minimize token footprints;…

cs.CLcs.AIcs.LG

KVEraser: Learning to Steer KV Cache for Efficient Localized Context Erasing [15]

Mufei Li, Shikun Liu, Dongqi Fu, Haoyu Wang, Yinglong Xia, Hong Li, Hong Yan, Pan Li

Post-hoc context erasing over the KV cache is challenging because a local edit has a global consequence: once a span has been processed, its influence propagates into the cached states of all…

cs.CLcs.LG

🧠 Privacy / Your / Cloak 2

Your Privacy My Cloak: Backdoor Attacks on Differentially Private Federated Learning [14]

Xiaolin Li, Ning Wang, Ninghui Li, Wenhai Sun

Prior research suggests that differential privacy (DP) inherently enhances the robustness of federated learning (FL) against backdoor attacks. In this paper, we challenge this assumption. Through an…

cs.LGcs.CR

Cross-Silo De-Anonymization Under Local Differential Privacy: Threat Model, Phase Transition, and Coordination Necessity [27]

Ziniu Liu, Aiping Li

When a person's records appear in k independent data silos, each protected by (epsilon, delta)-differential privacy, standard composition yields a valid (k*epsilon, k*delta)-DP guarantee for the…

cs.CRcs.ITcs.LG

🧠 Mpx / Unified / Systolic 2

FEnc$^2$: Unifying Data Packing for Efficient Private Inference via Convolution and Architecture-Aware Fragment Encoding [35]

Ran Ran, Zhaoting Gong, Nuo Xu, Yuanchao Xu, Fan Yao, Wujie Wen

Fully Homomorphic Encryption (FHE) enables privacy-preserving machine learning but incurs extreme computational and memory overhead. These costs come not only from expensive low-level primitives,…

cs.CRcs.LG

MPX: A Unified Systolic Array for Matrix and Polynomial Multiplication [34]

George Alexakis, Dimitrios Schoinianakis, Giorgos Dimitrakopoulos

Polynomial multiplication is a fundamental kernel in Fully Homomorphic Encryption (FHE) and post-quantum cryptography (PQC) and is commonly accelerated through Number Theoretic Transforms (NTTs). To…

cs.CRcs.AR

🧠 Importance / Phase / Representations 1

The Importance of Phase in Neural Representations: An Internal Oppenheim-Lim Test of Image Classifiers [1]

Alper Yıldırım

Oppenheim and Lim (1981) showed that natural images stay recognizable when reconstructed from their Fourier phase alone, while the magnitude carries little of their identity. We ask whether trained…

cs.CVcs.AIcs.LG

🧠 Hamon / Passive / Optical 1

HAMON: Passive Optical Sequence Mixing for Long-Horizon Forecasting [2]

Alper Yıldırım

Simple linear and frequency-domain models remain surprisingly competitive in long-horizon time-series forecasting, and recent mechanistic evidence suggests that standard forecasting benchmarks may…

cs.LGcs.AIcs.AR

🧠 Fusionrs / Large-scale / Rgb-infrared 1

FusionRS: A Large-Scale RGB-Infrared Remote Sensing Dataset for Dual-Modal Vision-Language Foundation Models [3]

Jiaju Han, Ben Zhang, Xuemeng Sun, Qike Zhang, Yuxian Dong, Chengyin Hu, Fengyu Zhang, Yiwei Wei, Jiujiang Guo

Remote sensing vision-language models have advanced Earth observation understanding, but most existing work remains centered on RGB imagery, leaving the complementary information in infrared data…

cs.CVcs.AI

🧠 Tunejury / Open / Metric 1

TuneJury: An Open Metric for Improving Music Generation Preference Alignment [5]

Yonghyun Kim, Junwon Lee, Haiwen Xia, Yinghao Ma, Junghyun Koo, Koichi Saito, Yuki Mitsufuji, Chris Donahue

We introduce TuneJury, an open, instance-level pairwise reward model for text-to-music that predicts a music preference score from a text prompt and an audio clip. The released checkpoint is trained…

cs.SDcs.AIcs.LG

🧠 Bayesian / Inference / Decision 1

Bayesian Inference and Decision Audits for Public Archives of Frontier AI Evaluations [6]

Yanan Long

Public AI evaluations are often read as terminal leaderboards, yet the underlying evidence is a selective time series shaped by reporting rules, benchmark revisions, and missingness. Repeated public…

cs.AIstat.ME

🧠 Activesam / Image-conditional / Class 1

ActiveSAM: Image-Conditional Class Pruning for Fast and Accurate Open-Vocabulary Segmentation [7]

Tran Dinh Tien, Zhiqiang Shen

Segment Anything Model 3 (SAM 3) provides a strong frozen backbone for concept-prompted segmentation, but applying it directly to open-vocabulary semantic segmentation (OVSS) is inefficient:…

cs.CVcs.AIcs.LG

🧠 Stable / Menus / Public 1

Stable Menus of Public Goods: AI-Enabled Progress [9]

Sara Fish

Using an open problem from the EC 2025 paper "Stable Menus of Public Goods" as a testbed, we conduct experiments to understand the effectiveness of different AI-for-EconCS research workflows.…

cs.GTcs.AIcs.CY

🧠 Consensus-based / Agentic / Framework 1

Consensus-based Agentic Large Language Model Framework for Harmonized Tariff Schedule Code Classification [10]

Truong Thanh Hung Nguyen, Khanh Van Quynh Nguyen, Hoang-Loc Cao, Tri Duong, Phuc Ho, Van Pham, Loc Nguyen, Hung Cao

Accurate Harmonized Tariff Schedule (HTS) code classification is essential for customs clearance, duty assessment, trade statistics, and regulatory compliance in maritime logistics. However, exact…

cs.AI

🧠 Exact / Posterior / Score 1

Exact Posterior Score Estimation for Solving Linear Inverse Problems [11]

Abbas Mammadov, Ozgur Kara, Kaan Oktay, Iskander Azangulov, Adil Kaan Akan, Hyungjin Chung, James Matthew Rehg, Yee Whye Teh

Diffusion and flow-based models learn powerful data priors by training a denoiser to reverse Gaussian corruption. To use this prior to solve a linear inverse problem, one needs to sample from the…

cs.LGcs.CVstat.ML

🧠 Hierarchical / Advantage / Weighting 1

Hierarchical Advantage Weighting for Online RL Fine-Tuning of VLAs from Sparse Episode Outcomes [13]

Tongyan Fang, Siyuan Huang, Naiyu Fang, Ganlong Zhao, Zhongjin Luo, Jianbo Liu, Xiaogang Wang, Ying Dong, Hongsheng Li

When pretrained VLA policies are fine-tuned through online RL, each rollout episode produces only a single binary outcome (success or failure), yet the actor update requires per-transition…

cs.ROcs.LG

🧠 Geometry / Data / Mathematical 1

Learning the Geometry of Data: A Mathematical Review of Shape Space Analysis [17]

Gary P. T. Choi, Khanh Dao Duc, Shira Faigenbaum-Golovin, Karen Habermann, Emmanuel Hartman, Christoph von Tycowicz, Chi Zhang, Wenjun Zhao, Felix Zhou

A central objective of machine learning is to identify structure and patterns in data. Advances in data acquisition have increasingly produced datasets whose observations possess rich geometric form,…

math.STcs.LGstat.ML

🧠 Brdfusion / Physics / Meets 1

BRDFusion: Physics Meets Generation for Urban Scene Inverse Rendering [22]

Yi-Ruei Liu, Jie-Ying Lee, Zheng-Hui Huang, Yu-Lun Liu, Chih-Hao Lin

Inverse rendering of urban scenes from captured videos enables numerous applications, including content creation and autonomous driving simulation. Physically-based rendering methods follow and…

cs.CV

🧠 Simultaneous / Tricolor / Video 1

Simultaneous Tricolor Video Observations of Three Tiny Near-Earth Asteroids with Sub-Minute Rotation Periods [23]

Jin Beniyama, Ryou Ohsawa, Shigeyuki Sako, Satoshi Takita, Daisuke Kuroda, Tomohiko Sekiguchi, Keisuke Isogai

Studying the physical properties of near-Earth asteroids (NEAs) is crucial for understanding their dynamical histories and origins, and assessing impact hazards to Earth. Tiny NEAs with diameters…

astro-ph.EPastro-ph.IM

🧠 Di5guise / Privacy / Vsim 1

Di5Guise: 5G Privacy with vSIM [24]

Shirin Ebadi, Zach Moolman, Eric Keller, Tamara Lehman

SIM cards have been the key building block of user authenticationand security in cellular networks. While they are meant to serve as privacy protecting elements in cellular communications, they can…

cs.CRcs.NI

🧠 Ghosts / Polymarket / When 1

The Ghosts of Polymarket: When Off-Chain Matches Meet On-Chain Reverts [25]

Yiming Shen, Yuhan Jin, Shuohan Wu, Yanlin Wang, Jiachi Chen

Polymarket has emerged as a prominent prediction market platform and one of the fastest-growing applications in DeFi. To achieve low-latency trading, it adopts a hybrid architecture that matches…

cs.CR

🧠 How / Much / Can 1

How Much Can We Trust LLM Search Agents? Measuring Endorsement Vulnerability to Web Content Manipulation [26]

Yimeng Chen, Zhe Ren, Firas Laakom, Yu Li, Dandan Guo, Jürgen Schmidhuber

Large language model (LLM)-based search agents synthesize open-web content into actionable recommendations on behalf of users, creating a risk that attacker-published pages are transformed into…

cs.CLcs.CRcs.CY

🧠 Automated / Jailbreak / Attack 1

Automated jailbreak attack targeting multiple defense strategies [28]

Qi Wang, Chengcheng Wan, Weijia He, Yanqing Li, Hanqi Sun, Xiaodong Gu, Jiangtao Wang

Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks. However, their safety remains a critical concern due to their susceptibility to adversarial…

cs.CRcs.AI

🧠 Robust / Automated / Reconfiguration 1

Robust and Automated Reconfiguration of Byzantine Wide-Area Replication [29]

Rowdy Chotkan, Bulat Nasrulin, Johan Pouwelse, Jérémie Decouchant

Distributed systems handle adversarial nodes through redundancy, which imposes a significant performance overhead. In blockchain systems, Byzantine fault-tolerant state-machine replication (BFT-SMR)…

cs.DCcs.CRcs.NI

🧠 Third-party / First-party / Measuring 1

From Third-Party to First-Party: Measuring and Protecting Against Modern Web Tracking Mechanisms [30]

Christian Böttger, Tareq Khouja, Norbert Pohlmann, Nurullah Demir, Tobias Urban

Web user tracking has always been a cat-and-mouse game between privacy-conscious users and trackers. Recently, this conflict has driven a shift from third-party tracking toward first-party tracking…

cs.CR

🧠 Jordan / Rigidity / Full 1

Jordan rigidity of full matrix algebras [31]

Ilja Gogić, Matija Kazalicki, Mateo Tomašević

Let $\mathbb{F}$ be a field of characteristic different from $2$, and let $M_n(\mathbb{F})^+$ denote the Jordan algebra of all $n\times n$ matrices over $\mathbb{F}$ with product $X\circ…

math.RA

🧠 Homomorphisms / Hook / Specht 1

Homomorphisms to Hook Specht Modules over Quiver Hecke Algebras of Type C [32]

Martín Forsberg Conde, Berta Hudak

We investigate homomorphisms between Specht modules for level 1 cyclotomic quiver Hecke algebras of affine type C. For hook partitions with a `short leg' or a `short arm', we give the full quiver…

math.RT

🧠 Measurement / Post-quantum / Readiness 1

Measurement Study of Post-Quantum Readiness of Internet: 2026 [33]

Vanishka Mohan Dubey, Gaurav Varshney

The emergence of quantum computing presents a fundamental challenge to the security of current Internet communication systems. Transport Layer Security (TLS), which forms the backbone of secure web…

cs.CR

🧠 Reconstruction / Time-dependent / Coefficients 1

Reconstruction of time-dependent coefficients in a semilinear dynamical Schr{ö}dinger equation [37]

Parveen Kumar, Gen Nakamura, Manmohan Vashisth

In the present manuscript, we study an inverse problem related to a semilinear dynamical Schr{ö}dinger equation with lower order terms, in a bounded domain of $\Rb^{1+n},n\geq 2$. Our focus is on…

math.AP

🧠 Complexity / Min-max / Optimization 1

The Complexity of Min-Max Optimization for Quadratic Polynomials [38]

Martino Bernasconi, Matteo Castiglioni, Andrea Celli, Alexandros Hollender

We prove that computing approximate stationary points of min-max optimization over the hypercube is PPAD-hard for quadratic polynomials. This holds even when the polynomials are multilinear, each…

cs.CCcs.GTcs.LG

🧠 Selection / Without / Signal 1

Selection Without Signal, Recovery Through Expression: A Measurement Study of Post-Hoc Falsification Operators for Frozen Small Code Models [39]

Mehmet Iscan

Frozen small code models (<=1.5B parameters, run locally without fine-tuning) suit offline and privacy-constrained use, but often emit plausible-but-wrong programs. A natural remedy is a post-hoc…

cs.SEcs.CLcs.LG

🧠 States / Compact / Spin-charge 1

Compact Spin-Charge Separated Neural Quantum States for Valence-Bond States [40]

Ang-Kun Wu, Louis Primeau, Yixin Zhang, Jingtao Zhang, Adrian Del Maestro, Yang Zhang

Neural-network quantum states (NQS) provide a flexible nonlinear representation of quantum many-body wavefunctions, but their efficiency depends sensitively on whether the architecture reflects the…

cond-mat.str-elcond-mat.dis-nn

🧠 Grid-state / Deformation / No-jump 1

Grid-state deformation in a no-jump non-Hermitian bosonic dimer [41]

B. M. Rodriguez-Lara, H. Ghaemi-Dizicheh, S. Dehdashti, A. Hanke, A. Touhami, J. Nötzel

We study the no-jump evolution of ideal grid states in a lossy bosonic dimer with differential decay. The effective non-Hermitian quadratic dynamics induces a complex symplectic flow in phase space…

quant-ph

🧠 Bath / Memory / Precision 1

Bath memory as a precision resource in quantum transport [42]

José Molina, Sheikh Parvez Mandal, Mahasweta Pandit, Javier Prior

Structured baths can reshape transport fluctuations in mesoscopic quantum devices, yet a predictive criterion for when this enhances precision has been lacking. We propose a route towards such…

quant-phcond-mat.mes-hallcond-mat.stat-mech

🧠 Exact / Value / Ramsey 1

The exact value of the Ramsey number $R(K_4-e,K_7)$ [43]

William J. Wesley

We compute the exact value of the Ramsey number $R(K_4-e,K_7)$. It is equal to 28.

math.CO

🧠 Hadronic / Tensor / Lattice 1

Hadronic tensor in lattice gauge theories by quantum computing [45]

Dairui Zou, Tianyin Li, Jian Liang, Enke Wang, Hongxi Xing

The hadronic tensor encodes crucial information regarding the internal structure of hadrons, reflecting the non-perturbative features of quantum chromodynamics (QCD). In this work, we directly…

hep-phhep-latnucl-th

🧠 Benchmarking / Llm / Agents 1

Benchmarking LLM Agents on Meta-Analysis Articles from Nature Portfolio [46]

Anzhe Xie, Weihang Su, Yujia Zhou, Yiqun Liu, Qingyao Ai

Meta-analysis is a demanding form of evidence synthesis that combines literature retrieval, PI/ECO-guided study selection, and statistical aggregation. Its structured, verifiable workflow makes it an…

cs.CLcs.IR

🧠 Distributed / Acoustic / Sensing 1

Distributed Acoustic Sensing for Urban Monitoring: Coverage Thresholds and Percolation [47]

Khen Cohen, Ariel Lellouch

Distributed Acoustic Sensing (DAS) enables the repurposing of existing fiber-optic networks as ultra-dense, long-range seismic arrays for urban monitoring. However, constraints imposed by real-world…

cond-mat.stat-mechphysics.app-phphysics.geo-ph

🧠 Filtered / Conformal / Ellipsoids 1

Filtered Conformal Ellipsoids for Graph-Native Time Series [48]

Yannick Limmer

Joint prediction sets for multivariate time series should control a single event while adapting to cross-coordinate dependence. We study filtered conformal ellipsoids: a frozen state-space filter…

cs.LGmath.STstat.ML

🧠 Constant-factor / Approximation / Gromov-hausdorff 1

A constant-factor approximation of the Gromov-Hausdorff distance in the plane [51]

Sushovan Majhi

We give the first polynomial-time constant-factor approximation of the Gromov--Hausdorff distance $d_{GH}$ between finite point sets in the Euclidean plane; in fixed Euclidean dimension such an…

cs.CGcs.DSmath.MG

🧠 Galaxy-cluster-stacked / Fermi-lat / Part 1

Galaxy-cluster-stacked Fermi-LAT, part IV: $\sim70$ GeV WIMP annihilation lines [52]

Uri Keshet

The strongest constraints on the velocity-dependent ($p$-wave) annihilation of weakly interacting massive particle (WIMP) dark matter were derived from the deep potential wells of galaxy clusters.…

astro-ph.HE

🌐 Web Findings

🤗 HuggingFace Papers 26

Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs [60]

Standard accuracy benchmarks are designed to test how closely large language models (LLMs) approach correct answers, but are not suitable for testing whether LLMs stick with a…

▲ 1HuggingFace Papers

VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models [69]

This technical report introduces VibeThinker-3B, a compact dense model with 3B parameters developed to investigate how far verifiable reasoning can be pushed within a strictly…

▲ 19HuggingFace Papers

Artificial Intelligence Index Report 2026 [65]

Welcome to the ninth edition of the AI Index report. As AI continues to advance rapidly, the question becomes whether the systems built around it can keep up. Governance…

▲ 3HuggingFace Papers

BadWorld: Adversarial Attacks on World Models [71]

Visual world models (VWMs) synthesize interactive, action-conditioned rollouts from a single context image. However, it remains an open question how robust these models are to…

▲ 13HuggingFace Papers

Who Should Lead Decoding Now? Tracking Reliable Trajectories for Ensembling Masked Diffusion Language Models [74]

Masked Diffusion Language Models (MDLMs) have emerged as a distinct paradigm for sequence generation. As MDLMs become diverse in capabilities and knowledge coverage, an important…

▲ 23HuggingFace Papers

JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence [67]

Many moments in the real world do not wait for a user to ask. A fire starts on a security monitor, an expression flickers across a video call, or a product a viewer wants flashes…

▲ 141HuggingFace Papers

StepPO: Step-Aligned Policy Optimization for Agentic Reinforcement Learning [78]

Agentic reinforcement learning (RL) is emerging as a critical post-training paradigm for improving LLM agent capabilities. Existing RL algorithms for LLMs largely follow the…

▲ 3HuggingFace Papers

CODA-BENCH: Can Code Agents Handle Data-Intensive Tasks? [75]

Advanced agents are increasingly demonstrating the potential to operate as autonomous engineers, creating a growing demand for evaluation benchmarks that capture the complexity of…

▲ 11HuggingFace Papers

Selective Control under Noisy Perception: Governance Failures Hidden by Aggregate Metrics in Modular Networks [59]

A content-moderation system can score well on every standard accuracy metric and still cause real harm, if its mistakes fall on the few users who connect otherwise separate…

▲ 1HuggingFace Papers

FastContext: Training Efficient Repository Explorer for Coding Agents [77]

Large Language Model (LLM) coding agents have achieved strong results on software engineering tasks, yet repository exploration remains a major bottleneck: locating relevant code…

▲ 40HuggingFace Papers

Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning [68]

We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20…

▲ 3HuggingFace Papers

Implicit Reasoning for Large Language Model-based Generative Recommendation [63]

Large Language Models (LLMs) are increasingly adopted as backbones for Generative Recommendation (GR), promising access to pretrained world knowledge. Yet reliably invoking this…

HuggingFace Papers

UniDDT: Unifying Multimodal Understanding and Generation with Decoupled Diffusion Transformer [76]

Unified Multimodal Models (UMMs) have emerged as a critical direction for general-purpose multimodal intelligence, integrating understanding and generation into a single…

▲ 7HuggingFace Papers

VisualClaw: A Real-Time, Personalized Agent for the Physical World [72]

Vision language models are serving as general-purpose interfaces for complex multimodal tasks. However, deployment still faces three gaps: VLMs typically incur high latency and…

▲ 18HuggingFace Papers

Where Did It Go Wrong? Process-Level Evaluation of Web Agents with Semantic State Tracking [61]

Web agents act through long interaction sequences, yet existing benchmarks evaluate only terminal success, discarding all process information and offering little guidance on…

▲ 10HuggingFace Papers

GD^2PO: Mitigating Multi-Reward Conflicts via Group-Dynamic reward-Decoupled Policy Optimization [56]

As LLMs advance, post-training reinforcement learning (RL) increasingly relies on multi-dimensional rewards to cultivate comprehensive capabilities. This shift demands new…

▲ 8HuggingFace Papers

Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test Time [73]

Extending a vision-language-action (VLA) policy to a new task typically requires task-specific teleoperated demonstrations and per-task fine-tuning, making adaptation costly in…

▲ 9HuggingFace Papers

Tangram: Unlocking Non-Uniform KV Cache Compression for Efficient Multi-turn LLM Serving [66]

Multi-turn LLM serving accumulates dialogue history whose Key-Value (KV) cache grows with every turn and every user, quickly exceeding the model weights themselves and making…

▲ 7HuggingFace Papers

DreamX-World 1.0: A General-Purpose Interactive World Model [58]

DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation. It supports camera navigation, revisits to previously…

▲ 60HuggingFace Papers

OneRank: Unified Transformer-Native Ranking Architecture for Multi-Task Recommendation [70]

Multi-task learning (MTL) is essential in recommender systems to enable complementary learning among diverse user feedback. While modern industrial practices have shifted from…

▲ 13HuggingFace Papers

Memento: Reconstruct to Remember for Consistent Long Video Generation [54]

Long-form video generation requires recurring subjects to remain consistent across various shots, viewpoints, motions, and scene transitions. Existing temporal decomposition…

▲ 7HuggingFace Papers

SP^3: Spherical Priors for Plug-and-Play Restoration [55]

In this paper, we introduce SP^3, a novel Plug-and-Play algorithm that accelerates maximum a posteriori image restoration by replacing denoisers with Spherical Encoders (SE) as…

▲ 3HuggingFace Papers

PhoneHarness: Harnessing Phone-Use Agents through Mixed GUI, CLI, and Tool Actions [64]

Phone agents are increasingly expected to complete real mobile workflows rather than merely predict the next screen action. However, much of the current mobile-agent literature…

▲ 7HuggingFace Papers

PermaVid: Consistent Video Generation Across Edits via Disentangled Context Memory [62]

Consistent video generation under editing operations requires persistence: when edits modify scene appearance or layout, subsequent generations should remain coherent across time…

▲ 1HuggingFace Papers

MMDiff: Extending Diffusion Transformers for Multi-Modal Generation [57]

Diffusion transformers have demonstrated remarkable generative capabilities, yet the rich perceptual representations computed across their denoising trajectory are discarded once…

▲ 1HuggingFace Papers

Steady-Forcing: Balancing Spatial Persistence and Motion Continuity in Long-Horizon Nature Video Diffusion [79]

Autoregressive video diffusion models enable streaming generation but often degrade over long rollouts: static scene layouts drift, while mechanisms that improve spatial stability…

HuggingFace Papers

📝 OpenReview 2

SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression [89]

The advancements in Large Language Models (LLMs) have been hindered by their substantial sizes, which necessitate LLM compression methods for practical deployment. Singular Value…

OpenReview

Diffusion$^2$: Dynamic 3D Content Generation via Score Composition of Video and Multi-view Diffusion Models [90]

Recent advancements in 3D generation are predominantly propelled by improvements in 3D-aware image diffusion models. These models are pretrained on Internet-scale image data and…

OpenReview

🦞 Lobste.rs 1

Why adding ontologies to LLMs won't yield machine intelligence [53]

Finding from 🦞 Lobste.rs.

Lobste.rs