Search: “inference”

MemGround: Long-Term Memory Evaluation Kit for Large Language Models in Gamified Scenarios

model-releasesarxiv-cs-cl

arXiv:2604.14158v1 Announce Type: new Abstract: Current evaluations of long-term memory in LLMs are fundamentally static. By fixating on simple retrieval and short-context inference, they neglect the

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Research

No More Guessing: a Verifiable Gradient Inversion Attack in Federated Learning

DGX agent

arXiv:2604.15063v1 Announce Type: new Abstract: Gradient inversion attacks threaten client privacy in federated learning by reconstructing training samples from clients' shared gradients. Gradients ag

researcharxiv-cs-lg

17 Apr 2026

PAGE-4D: Disentangled Pose and Geometry Estimation for VGGT-4D Perception

arXiv:2510.17568v5 Announce Type: replace Abstract: Recent 3D feed-forward models, such as the Visual Geometry Grounded Transformer (VGGT), have shown strong capability in inferring 3D attributes of s

applicationsarxiv-cs-cv

Route to Rome Attack: Directing LLM Routers to Expensive Models via Adversarial Suffix Optimization

arXiv:2604.15022v1 Announce Type: cross Abstract: Cost-aware routing dynamically dispatches user queries to models of varying capability to balance performance and inference cost. However, the routing

applicationsarxiv-cs-cl

StreamCacheVGGT: Streaming Visual Geometry Transformers with Robust Scoring and Hybrid Cache Compression

arXiv:2604.15237v1 Announce Type: new Abstract: Reconstructing dense 3D geometry from continuous video streams requires stable inference under a constant memory budget. Existing O(1) frameworks primar

researcharxiv-cs-cv

Beyond State Consistency: Behavior Consistency in Text-Based World Models

arXiv:2604.13824v1 Announce Type: new Abstract: World models have been emerging as critical components for assessing the consequences of actions generated by interactive agents in online planning and

safetyarxiv-cs-lg

Hardware

Event Tensor: A Unified Abstraction for Compiling Dynamic Megakernel

arXiv:2604.13327v1 Announce Type: cross Abstract: Modern GPU workloads, especially large language model (LLM) inference, suffer from kernel launch overheads and coarse synchronization that limit inter

hardwarearxiv-cs-lg

Free Geometry: Refining 3D Reconstruction from Longer Versions of Itself

model-releasesarxiv-cs-cv

arXiv:2604.14048v1 Announce Type: new Abstract: Feed-forward 3D reconstruction models are efficient but rigid: once trained, they perform inference in a zero-shot manner and cannot adapt to the test s

From Where Words Come: Efficient Regularization of Code Tokenizers Through Source Attribution

arXiv:2604.14053v1 Announce Type: new Abstract: Efficiency and safety of Large Language Models (LLMs), among other factors, rely on the quality of tokenization. A good tokenizer not only improves infe

safetyarxiv-cs-cl

GeoBridge: A Semantic-Anchored Multi-View Foundation Model Bridging Images and Text for Geo-Localization

model-releasesarxiv-cs-cv

arXiv:2512.02697v3 Announce Type: replace Abstract: Cross-view geo-localization infers a location by retrieving geo-tagged reference images that visually correspond to a query image. However, the trad

Reward Design for Physical Reasoning in Vision-Language Models

model-releasesarxiv-cs-cl

arXiv:2604.13993v1 Announce Type: cross Abstract: Physical reasoning over visual inputs demands tight integration of visual perception, domain knowledge, and multi-step symbolic inference. Yet even st

Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model

arXiv:2510.18165v2 Announce Type: replace-cross Abstract: Diffusion language models (DLMs) are emerging as a powerful and promising alternative to the dominant autoregressive paradigm, offering inhere

Towards Unconstrained Human-Object Interaction

arXiv:2604.14069v1 Announce Type: new Abstract: Human-Object Interaction (HOI) detection is a longstanding computer vision problem concerned with predicting the interaction between humans and objects.

researcharxiv-cs-cv

Agents

Aethon: A Reference-Based Replication Primitive for Constant-Time Instantiation of Stateful AI Agents

arXiv:2604.12129v1 Announce Type: new Abstract: The transition from stateless model inference to stateful agentic execution is reshaping the systems assumptions underlying modern AI infrastructure. Wh

agentsarxiv-cs-ai

CamReasoner: Reinforcing Camera Movement Understanding via Structured Spatial Reasoning

arXiv:2602.00181v3 Announce Type: replace-cross Abstract: Understanding camera dynamics is a fundamental pillar of video spatial intelligence. However, existing multimodal models predominantly treat t

Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation

arXiv:2604.12424v1 Announce Type: cross Abstract: Multimodal Large Language Models frequently suffer from inference hallucinations, partially stemming from language priors dominating visual evidence.

Does Visual Token Pruning Improve Calibration? An Empirical Study on Confidence in MLLMs

arXiv:2604.12035v1 Announce Type: new Abstract: Visual token pruning is a widely used strategy for efficient inference in multimodal large language models (MLLMs), but existing work mainly evaluates i

researcharxiv-cs-cv

Enhance-then-Balance Modality Collaboration for Robust Multimodal Sentiment Analysis

arXiv:2604.12518v1 Announce Type: new Abstract: Multimodal sentiment analysis (MSA) integrates heterogeneous text, audio, and visual signals to infer human emotions. While recent approaches leverage c

Hybrid-AIRL: Enhancing Inverse Reinforcement Learning with Supervised Expert Guidance

arXiv:2511.21356v2 Announce Type: replace-cross Abstract: Adversarial Inverse Reinforcement Learning (AIRL) has shown promise in addressing the sparse reward problem in reinforcement learning (RL) by

Information-Theoretic Optimization for Task-Adapted Compressed Sensing Magnetic Resonance Imaging

arXiv:2604.12709v1 Announce Type: cross Abstract: Task-adapted compressed sensing magnetic resonance imaging (CS-MRI) is emerging to address the specific demands of downstream clinical tasks with sign

Latent-Condensed Transformer for Efficient Long Context Modeling

arXiv:2604.12452v1 Announce Type: new Abstract: Large language models (LLMs) face significant challenges in processing long contexts due to the linear growth of the key-value (KV) cache and quadratic

Not All Turns Are Equally Hard: Adaptive Thinking Budgets For Efficient Multi-Turn Reasoning

arXiv:2604.05164v2 Announce Type: replace-cross Abstract: As LLM reasoning performance plateau, improving inference-time compute efficiency is crucial to mitigate overthinking and long thinking traces

Observing the unobserved confounding through its effects: toward randomized trial-like estimates from real-world survival data

arXiv:2604.12137v1 Announce Type: cross Abstract: Background: Randomized controlled trials (RCTs) are costly, time-consuming, and often infeasible, while treatment-effect estimation from observational

applicationsarxiv-cs-ai

OSC: Hardware Efficient W4A4 Quantization via Outlier Separation in Channel Dimension

arXiv:2604.12782v1 Announce Type: cross Abstract: While 4-bit quantization is essential for high-throughput deployment of Large Language Models, activation outliers often lead to significant accuracy

SpecBranch: Speculative Decoding via Hybrid Drafting and Rollback-Aware Branch Parallelism

arXiv:2506.01979v4 Announce Type: replace-cross Abstract: Recently, speculative decoding (SD) has emerged as a promising technique to accelerate LLM inference by employing a small draft model to propo

applicationsarxiv-cs-ai

Absorption and Inertness in Coarse-Grained Arithmetic: A Heuristic Application to the St. Petersburg Paradox

arXiv:2507.12475v2 Announce Type: replace-cross Abstract: The St. Petersburg paradox presents a longstanding challenge in decision theory: its classical expected value diverges, yet no correspondingly

CID-TKG: Collaborative Historical Invariance and Evolutionary Dynamics Learning for Temporal Knowledge Graph Reasoning

arXiv:2604.09600v1 Announce Type: new Abstract: Temporal knowledge graph (TKG) reasoning aims to infer future facts at unseen timestamps from temporally evolving entities and relations. Despite recent

CLASP: Closed-loop Asynchronous Spatial Perception for Open-vocabulary Desktop Object Grasping

arXiv:2604.11320v1 Announce Type: new Abstract: Robot grasping of desktop object is widely used in intelligent manufacturing, logistics, and agriculture.Although vision-language models (VLMs) show str

applicationsarxiv-cs-ro

COREY: A Prototype Study of Entropy-Guided Operator Fusion with Hadamard Reparameterization for Selective State Space Models

arXiv:2604.10597v1 Announce Type: cross Abstract: State Space Models (SSMs), represented by the Mamba family, provide linear-time sequence modeling and are attractive for long-context inference. Yet p

Defending against Backdoor Attacks via Module Switching

arXiv:2504.05902v2 Announce Type: replace-cross Abstract: Backdoor attacks pose a serious threat to deep neural networks (DNNs), allowing adversaries to implant triggers for hidden behaviors in infere

ECHO: Elastic Speculative Decoding with Sparse Gating for High-Concurrency Scenarios

arXiv:2604.09603v1 Announce Type: cross Abstract: Speculative Decoding promises to accelerate the inference of Large Language Models, yet its efficacy often degrades in production-grade serving. Exist

applicationsarxiv-cs-ai

Exact Certification of Neural Networks and Partition Aggregation Ensembles against Label Poisoning

arXiv:2604.11416v1 Announce Type: new Abstract: Label-flipping attacks, which corrupt training labels to induce misclassifications at inference, remain a major threat to supervised learning models. Th

researcharxiv-cs-lg

F2F-AP: Flow-to-Future Asynchronous Policy for Real-time Dynamic Manipulation

arXiv:2604.02408v2 Announce Type: replace Abstract: Asynchronous inference has emerged as a prevalent paradigm in robotic manipulation, achieving significant progress in ensuring trajectory smoothness

safetyarxiv-cs-ro

FRAMER: Frequency-Aligned Self-Distillation with Adaptive Modulation Leveraging Diffusion Priors for Real-World Image Super-Resolution

arXiv:2512.01390v3 Announce Type: replace Abstract: Real-image super-resolution (Real-ISR) seeks to recover HR images from LR inputs with mixed, unknown degradations. While diffusion models surpass GA

safetyarxiv-cs-cv

HDR Video Generation via Latent Alignment with Logarithmic Encoding

arXiv:2604.11788v1 Announce Type: new Abstract: High dynamic range (HDR) imagery offers a rich and faithful representation of scene radiance, but remains challenging for generative models due to its m

safetyarxiv-cs-cv

Layerwise Dynamics for In-Context Classification in Transformers

arXiv:2604.11613v1 Announce Type: cross Abstract: Transformers can perform in-context classification from a few labeled examples, yet the inference-time algorithm remains opaque. We study multi-class

Local Ai

Look Twice before You Leap: A Rational Framework for Localized Adversarial Anonymization

arXiv:2512.06713v3 Announce Type: replace-cross Abstract: Current LLM-based frameworks for text anonymization usually rely on remote API services from powerful LLMs, which creates an inherent privacy

local-aiarxiv-cs-cl

MemDLM: Memory-Enhanced DLM Training

model-releasesarxiv-cs-cl

arXiv:2603.22241v2 Announce Type: replace Abstract: Diffusion Language Models (DLMs) offer attractive advantages over Auto-Regressive (AR) models, such as full-attention parallel decoding and flexible

MonoEM-GS: Monocular Expectation-Maximization Gaussian Splatting SLAM

arXiv:2604.10593v1 Announce Type: new Abstract: Feed-forward geometric foundation models can infer dense point clouds and camera motion directly from RGB streams, providing priors for monocular SLAM.

safetyarxiv-cs-ro

Agents

MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens

arXiv:2603.23516v2 Announce Type: replace-cross Abstract: Long-term memory is a cornerstone of human intelligence. Enabling AI to process lifetime-scale information remains a long-standing pursuit in

agentsarxiv-cs-ai

Multidata Causal Discovery for Statistical Hurricane Intensity Forecasting

arXiv:2510.02050v3 Announce Type: replace-cross Abstract: Improving statistical forecasts of tropical cyclone (TC) intensity is limited by complex nonlinear interactions and difficulty in identifying

researcharxiv-cs-lg

Pay Less Attention to Function Words for Free Robustness of Vision-Language Models

arXiv:2512.07222v3 Announce Type: replace-cross Abstract: To address the trade-off between robustness and performance for robust VLM, we observe that function words could incur vulnerability of VLMs a

Prompt Injection as Role Confusion

arXiv:2603.12277v3 Announce Type: replace-cross Abstract: Language models remain vulnerable to prompt injection attacks despite extensive safety training. We trace this failure to role confusion: mode

Quantization Dominates Rank Reduction for KV-Cache Compression

model-releasesarxiv-cs-ai

arXiv:2604.11501v1 Announce Type: cross Abstract: We compare two strategies for compressing the KV cache in transformer inference: rank reduction (discard dimensions) and quantization (keep all dimens

Rebooting Microreboot: Architectural Support for Safe, Parallel Recovery in Microservice Systems

arXiv:2604.09963v1 Announce Type: cross Abstract: Microreboot enables fast recovery by restarting only the failing component, but in modern microservices naive restarts are unsafe: dense dependencies

SafeConstellations: Mitigating Over-Refusals in LLMs Through Task-Aware Representation Steering

arXiv:2508.11290v3 Announce Type: replace Abstract: LLMs increasingly exhibit over-refusal behavior, where safety mechanisms cause models to reject benign instructions that seemingly resemble harmful

safetyarxiv-cs-cl

Spatial Competence Benchmark

model-releasesarxiv-cs-ai

arXiv:2604.09594v1 Announce Type: new Abstract: Spatial competence is the quality of maintaining a consistent internal representation of an environment and using it to infer discrete structure and pla

Teaching Robots to Interpret Social Interactions through Lexically-guided Dynamic Graph Learning

arXiv:2604.10895v1 Announce Type: cross Abstract: For a robot to be called socially intelligent, it must be able to infer users internal states from their current behaviour, predict the users future b

researcharxiv-cs-ro