Search: “benchmark”

Srijika: OpenType-Layout-Reusing Font Restyling for Nine Indic Scripts

arXiv:2609.05661v1 Announce Type: new Abstract: We present Srijika, a system for producing installable OpenType fonts for nine Brahmic scripts: Devanagari, Tamil, Bengali, Telugu, Kannada, Malayalam,

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Research

STAR-Pro: Stage-Wise Token Adaptive Reduction with Progressive Refinement for Efficient Large Vision-Language Models

DGX agent

arXiv:2609.05916v1 Announce Type: new Abstract: Large vision-language models (LVLMs) achieve strong multimodal understanding, but the hundreds to thousands of visual tokens they process impose substan

researcharxiv-cs-cv

9 Sept 2026

Steering Geometry: Validating Human Value Geometry in LLM Steering Space

arXiv:2609.06289v1 Announce Type: cross Abstract: As large language models (LLMs) are increasingly deployed in alignment-sensitive contexts, activation steering has emerged as a lightweight, inference

Applications

Stochastically Perturbed Weights: Ensembles from Deterministic Machine-Learning Weather Models

arXiv:2609.08412v1 Announce Type: new Abstract: Machine-learning weather models (MLWMs) now match or outperform operational numerical weather prediction (NWP) at global medium-range forecasting, at fa

applicationsarxiv-cs-lg

Strategic Doctrine Language Models (sdLM): A Learning-System Framework for Doctrinal Consistency and Geopolitical Forecasting

arXiv:2601.14862v2 Announce Type: replace Abstract: We introduce Strategic Doctrine Language Models (sdLM), a learning-system framework for multi-document strategic reasoning with doctrinal consistenc

researcharxiv-cs-lg

Suan: Rectifying Direct Preference Safety Alignment in Large Language Models

arXiv:2609.08634v1 Announce Type: new Abstract: Integrating robust safety guardrails into Large Language Models (LLMs) is essential for delivering helpful yet harmless responses. While proprietary sys

Subspace Optimization for Backpropagation-Free Continual Test-Time Adaptation

arXiv:2603.28678v2 Announce Type: replace Abstract: We introduce PACE, a backpropagation-free continual test-time adaptation system that directly optimizes the affine parameters of normalization layer

researcharxiv-cs-lg

SUN: Reaching for Novelty in Reinforcement Learning

arXiv:2609.08642v1 Announce Type: new Abstract: Exploration in reinforcement learning (RL) remains a fundamental challenge. Recent goal-conditioned RL strategies (which select goals to encourage broad

Sustained Performance and Energy Accounting for Nonlinear Forecasting Across Classical and Simulated Quantum Models

arXiv:2510.25183v2 Announce Type: replace-cross Abstract: Energy-efficient AI should be evaluated across the full application pipeline, not only by lowest error or shortest training time. We study thi

TASG-Explore: Traversability-Aware Sector-Guided Exploration for Ground Robot on Uneven Terrain

model-releasesarxiv-cs-ro

arXiv:2609.08512v1 Announce Type: new Abstract: Autonomous exploration on uneven terrain requires ground robots to balance exploration efficiency, coverage completeness, and terrain safety. Detailed t

Local Ai

Task-driven Processing with Coarse-to-Fine Glimpse-based Active Perception

arXiv:2609.09025v1 Announce Type: new Abstract: State-of-the-art vision models process images in their entirety, lacking the ability to selectively zoom in on relevant regions. This limitation is part

local-aiarxiv-cs-cv

TDDN: Text-aligned Diffused DINO Network for Puzzle Understanding

arXiv:2609.07937v1 Announce Type: new Abstract: Structured visual reasoning, such as image puzzles, demands fine-grained visual perception, an ability current Vision Language Models (VLMs) lack. VLMs

safetyarxiv-cs-cv

Temporal Consistency Improves Generalization in Contextual Offline Meta Reinforcement Learning

arXiv:2603.02935v2 Announce Type: replace Abstract: Offline meta-reinforcement learning seeks to learn a policy that generalizes to new related tasks online. Context-based methods infer a task represe

TFTrack: A Template-Free Framework for Efficient 3D Point Cloud Tracking

arXiv:2609.07738v1 Announce Type: new Abstract: LiDAR-based 3D Single Object Tracking (3D SOT) is critical for robotic perception and navigation and aims to localize dynamic objects across frames in s

safetyarxiv-cs-cv

The BatchNorm Illusion: Diagnosing Normalization Artifacts in Machine Unlearning Evaluation

arXiv:2609.08901v1 Announce Type: new Abstract: Approximate machine unlearning aims to remove the influence of specific training data from a trained model without retraining from scratch. We identify

The Price of Consistency: Exploiting Visual Anchors for Multimodal Jailbreaking in Video Generation

arXiv:2609.07216v1 Announce Type: new Abstract: The rapid evolution of video generation has shifted the paradigm from pure text-driven to multi-conditional controllable generation, with reference imag

TontaubeV1: Streaming Text-to-Speech with Hierarchical Codec Modeling and Bounded Context

arXiv:2609.08703v1 Announce Type: cross Abstract: Text-to-speech systems often face a trade-off between natural prosody and efficient inference: higher perceptual quality typically comes at increased

Toward Sustainable Distributed LLM Inference: A Systems Synthesis and Research Agenda for an Energy-, Carbon-, and Cache-Aware llm-d Control Plane

arXiv:2609.05565v1 Announce Type: cross Abstract: Large language model (LLM) sustainability is increasingly a serving-systems problem, not only a training problem. In production, energy and carbon imp

Towards Bridging the Gap Between Offline and Iterative Alignment via Preference Distillation

arXiv:2609.06893v1 Announce Type: cross Abstract: Direct preference optimization DPO is a promising offline approach for aligning large language models (LLMs) due to its simplicity, computational effi

Tracking the Moving Frontier: Long-Short Term Advantage Estimator

arXiv:2609.06671v1 Announce Type: new Abstract: Group-based RLVR methods estimate advantages by repeatedly sampling multiple trajectories for each prompt, making long-horizon agent training expensive

TransNormal-2: Geometry-Grounded Rectified Flow with Edge-Aware Decoding for Precise Normal Estimation

arXiv:2609.06665v1 Announce Type: new Abstract: Diffusion-based models enable monocular geometry estimation, yet their pixel-space precision is limited by a shared, under-studied error source: VAE rec

Agents

TrojanWorld: Backdooring World-Model Agents via Imagination Steering

arXiv:2609.07051v1 Announce Type: new Abstract: World models increasingly serve as the predictive core of model-based reinforcement learning agents, enabling them to simulate future dynamics and reaso

agentsarxiv-cs-lg

Trust-But-Verify: Poisoning-Resilient Locally Private Graph Learning Protocols

arXiv:2609.07063v1 Announce Type: new Abstract: Built upon local differential privacy (LDP), locally private graph learning protocols have emerged as an important paradigm for decentralized graph lear

Two-Scale Localized PCA-Net: Coarse-Global and Local-Residual Representations for Artifact-Reduced PDE Operator Learning

arXiv:2609.08034v1 Announce Type: new Abstract: Localized dimensionality reduction improves the scalability of operator learning for high-dimensional partial differential equations (PDEs), but indepen

researcharxiv-cs-lg

Applications

Unified Multi-Layer Subspace Modeling for Cross-Domain OOD Detection

arXiv:2609.06785v1 Announce Type: new Abstract: Out-of-Distribution (OOD) detection remains a fundamental challenge for neural networks, whose predictions can be overconfident on inputs that deviate f

applicationsarxiv-cs-cv

US-JEPA: A Joint Embedding Predictive Architecture for Ultrasound

arXiv:2602.19322v2 Announce Type: replace Abstract: Ultrasound (US) imaging poses unique challenges for representation learning due to its inherently noisy acquisition process. The low signal-to-noise

VidaForge: Open Research Infrastructure for Video Pretraining Data Recipes

arXiv:2609.06652v1 Announce Type: new Abstract: Video foundation models increasingly rely on large-scale pretraining data, yet the end-to-end data pipelines behind them remain largely closed and diffi

Visible-Reachable Workspace for Perception-Aware Humanoid Design

model-releasesarxiv-cs-ro

arXiv:2609.08905v1 Announce Type: new Abstract: Workspace analysis measures where a robot can place its end effector. For visually guided manipulation, reachability alone is insufficient: a kinematica

Vision-Guided Text Prompt Tuning for Multimodal Sentiment Analysis

arXiv:2609.06497v1 Announce Type: cross Abstract: Multimodal sentiment analysis requires effective modeling of both verbal semantics and non-verbal affective cues. A central challenge is to calibrate

Applications

Voxeland: Probabilistic Instance-Aware Semantic Mapping with Evidence-based Uncertainty Quantification

arXiv:2411.08727v2 Announce Type: replace Abstract: Robots in human-centered environments require accurate scene understanding to perform high-level tasks effectively. This understanding can be achiev

applicationsarxiv-cs-ro

When Semantically Consistent Encoding Meets View-Label Heterogeneity Modeling: A Unified Framework for Incomplete Multi-View Multi-Label Learning

arXiv:2609.07525v1 Announce Type: new Abstract: Incomplete multi-view multi-label learning requires not only robust semantic aggregation from partially observed views, but also label-aware exploitatio

researcharxiv-cs-cv

Why shared attention vectors fail: a case for outcome-indexed tuning

arXiv:2609.08615v1 Announce Type: new Abstract: Dimensional attention in learning is often implemented as a globally shared attention vector, where each stimulus dimension corresponds to a single scal

ZetaDial: dialing net charge of protein binders at inference time for therapeutic developability

arXiv:2609.05451v1 Announce Type: cross Abstract: Net charge is a developability-relevant property of therapeutic binders, linked to viscosity, clearance, nonspecific interaction and aggregation, and

How KDDI built Buffmee, a faster, reliable consumer RAG app

model-releasesgoogle-cloud-ai

When building consumer-facing generative AI applications, balancing high generation quality with fast response times across diverse media types, can be challenging. KDDI, a major telecommunications ca

I made a custom llama.cpp build optimized for 7900xtx (one or two). for qwen 3.8 next and 27B. includes optimizations for PciE x4 and tensor parallel. read inside! (no AI slop)

model-releasesr-localllama

I found a lot of room on the table for these cards so I decided to make a specialized build to squeeze all I could. first The results: qwen 3.8 next Q3_K_XL: 920tk/s pp8192 (2 cards, ram offload), 24/

Nice report on agents beyond code generation. Here is why it matters: Coding agents raise how much code gets written. This report argues the…

Nice report on agents beyond code generation. Here is why it matters: Coding agents raise how much code gets written. This report argues the gains shrink sharply between writing code and shipping reli

model-releasesdair-ai--x

Local Ai

OpenAI alleged of stealing mathematicians work

Privacy have been concern of many of us to have their own hardware to run llms, and here's another reason why: two mathematicians spent a year cracking one of the hardest problems in math and fed ever

local-air-localllama

Pathway’s brain-inspired architecture development on Amazon SageMaker HyperPod

model-releasesaws-ml-blog

Pathway's Baby Dragon Hatchling (BDH) is a brain-inspired, post-transformer architecture that reasons in latent space instead of emitting chain-of-thought tokens. See how Pathway develops and scales B

practically my entire feed for days had consisted of string of reports about OpenAI’s misconduct, around the hacks, around apparent coverups…

model-releasesgary-marcus--x

practically my entire feed for days had consisted of string of reports about OpenAI’s misconduct, around the hacks, around apparent coverups, possible data fudging and maybe even intellectual theft wi

Industry

“The Earth is the cradle of humanity, but one cannot live in the cradle forever.” — Konstantin Tsiolkovsky The unique character of AI in Tes…

“The Earth is the cradle of humanity, but one cannot live in the cradle forever.” — Konstantin Tsiolkovsky The unique character of AI in Tesla and SpaceX is that same spirit of exploration: the true f

industryelon-musk--x

Local Ai

Which local model is actually good at knowing when to stop and ask you a question?

I’ve been thinking about this after using more agentic/local coding models. A lot of the newer models are surprisingly good at continuing on their own. But sometimes that seems like the problem. If a

local-air-localllama

A couple years later, with GPT-6. https://underdark-duel.netlify.app Apparently the mindflayer has a '101-bone skeletal rig' (This is 5E 201…

model-releasesethan-mollick--x

A couple years later, with GPT-6. https://underdark-duel.netlify.app Apparently the mindflayer has a '101-bone skeletal rig' (This is 5E 2014 rules, and I didn't see major errors, though the mind flay

A Fairness Audit of the Duckworth-Lewis-Stern Method: Format-Specific and Gender-Differential Bias, with an Interpretable Calibration Layer for Cricket Target Revision

arXiv:2609.04754v1 Announce Type: new Abstract: The Duckworth-Lewis-Stern (DLS) method has been the international standard for revising target scores in rain-interrupted limited-overs cricket since 19

A Roadmap for MEG Foundation Models

arXiv:2609.04461v1 Announce Type: cross Abstract: Foundation models are beginning to reshape brain-signal analysis by moving the field beyond task-specific decoding pipelines toward reusable models pr

researcharxiv-cs-ai

A Semantic Model of Genetic Evidence: A Step Toward Bridging the Basic-Science-Clinic Gap

model-releasesarxiv-cs-ai

arXiv:2609.04509v1 Announce Type: cross Abstract: Scientific and clinical decision-making depends on evidence from the primary literature, but existing standards for representing that evidence (FHIR E

A Tree-based RAG Framework for Evidence-Intensive QA via Adaptive Planning and Topology-Aware Evidence Gathering

arXiv:2609.04981v1 Announce Type: new Abstract: Recent structured RAG methods leverage tree- or graph-based reasoning structures to improve multi-hop QA. However, they face key limitations in evidence

researcharxiv-cs-ai

Agents

Abstraction Agent

arXiv:2609.04303v1 Announce Type: cross Abstract: Information abstraction, which groups strategically similar private states into a tractable number of buckets, is essential for scaling game-solving a

agentsarxiv-cs-ai

ACE: Adaptive Calibration-Free Expert Skipping for MoE-based LLMs

arXiv:2609.05228v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures provide an efficient paradigm for scaling large language models (LLMs), yet fixed top-k routing activates the sam

researcharxiv-cs-ai