4392 papers
The paper introduces Mozart, an algorithm-hardware co-design framework that leverages 3.5D wafer-scale chiplet architectures with specialized expert allocation and scheduling strategies to overcome communication and memory bottlenecks in the efficient training of large-scale Mixture-of-Experts (MoE) language models.
This study analyzes how 80 student design teams integrated generative AI into their creative process, revealing that while AI serves as a cognitive accelerator for early-stage tasks like brainstorming, human competencies in agency, domain knowledge, imagination, and taste remain essential for interpreting context, validating outputs, and refining design solutions.
This paper presents OV-DEIM, a real-time end-to-end DETR-style open-vocabulary object detector that combines the DEIMv2 framework with a query supplement strategy and a novel GridSynthetic data augmentation technique to achieve state-of-the-art performance and efficiency, particularly for rare categories.
This paper proposes the Hierarchical Memory Tree (HMT), a structured framework that decouples high-level task logic from site-specific action details through a three-level abstraction hierarchy, thereby significantly enhancing the generalization and robustness of large language model-based web agents in unseen environments.
This paper introduces TFM, a temporal attack framework that exploits the vulnerability of text-to-video models to generate harmful content by providing only sparse boundary conditions (start and end frames) and implicitly substituting sensitive cues, thereby bypassing existing safety filters and significantly increasing jailbreak success rates.
This paper demonstrates that leveraging eBPF-based system-level monitoring reveals new leakage patterns in Searchable Symmetric Encryption (SSE) that extend beyond traditional threat models, thereby enabling more powerful leakage abuse attacks and highlighting the critical need to address system-level exposures in SSE defenses.
This paper introduces the Safety-guaranteed Surgical Policy (SSP) framework, which integrates Neural ODE-based uncertainty modeling with robust Control Barrier Functions to enforce behavioral and spatial constraints, thereby ensuring near-zero safety violations while maintaining high task success rates in data-driven robot-assisted surgery.
This paper analyzes 377 YouTube videos to map the collective knowledge creators share about monetizing Generative AI, identifying ten common use cases and revenue strategies while highlighting structural tensions such as unverifiable income claims and shifting authorship norms in AI-mediated creative labor.
The paper introduces DeepEarth, a self-supervised multi-modal world model featuring Earth4D, a novel 4D space-time positional encoder that achieves state-of-the-art ecological forecasting performance and outperforms larger foundation models through efficient planetary-scale learning.
TacDexGrasp is a robust dexterous grasping framework that utilizes tactile feedback and a Second-Order Cone Programming controller to actively constrain tangential-to-normal force ratios, thereby preventing both translational and rotational slip without requiring explicit torque modeling or slip detection.
The paper introduces AIReSim, a discrete event simulator designed to help system designers evaluate and tune reliability mechanisms, prioritize improvements, and plan capacity for large-scale AI clusters by modeling the complex tradeoffs involved in failure, recovery, scheduling, and repair processes.
This paper proposes a novel fine-grained 3D facial reconstruction method for micro-expressions that integrates global dynamic features with locally-enriched cues from 2D motions, facial priors, and 3D geometry to overcome data scarcity and achieve superior geometric accuracy and perceptual detail compared to state-of-the-art approaches.
This paper investigates the potential of retrofitting to enhance sensor-powered smart car cabins by identifying limitations in built-in systems through interviews and defining user requirements via participatory design, ultimately offering design recommendations for future retrofit solutions.
This paper proposes CAPL, a framework that mitigates multi-image hallucinations in large vision-language models by introducing a selectable image token interaction mechanism for fine-grained cross-image alignment and a preference learning strategy that trains the model to rely on genuine visual evidence rather than textual priors.
This paper proposes a communication network-aware framework that integrates routing constraints with low-rank matrix completion to mitigate spatially correlated data losses and significantly improve missing data recovery accuracy in power distribution grids compared to traditional measurement-only approaches.
This paper introduces a scalable, domain-agnostic web-based framework that leverages Large Language Models to automate the collection, filtering, and construction of open scientific databases, achieving 90% overlap with expert-curated datasets while significantly reducing manual workload.
This paper presents a user-friendly framework that enables domain scientists to generate 3D animations of petascale, time-varying climate data on commodity hardware using an LLM-assisted conversational interface, thereby eliminating the need for specialized visualization expertise and high-performance computing resources.
This paper proposes a bi-directional digital twin prototype anchoring framework enhanced with multi-periodicity learning to achieve robust few-shot fault diagnosis by leveraging meta-training in a virtual simulation space and test-time adaptation in the physical domain, thereby overcoming the limitations of traditional methods that require abundant labeled or unlabeled target data.
SODA introduces a sensitivity-oriented dynamic acceleration framework for Diffusion Transformers that adaptively optimizes caching and pruning strategies through fine-grained sensitivity modeling and dynamic programming, achieving state-of-the-art generation fidelity under controllable acceleration ratios.
This paper introduces GuideTWSI, a diverse dataset combining synthetic and real-world images to address the scarcity of Tactile Walking Surface Indicator (TWSI) data, specifically bridging the gap between East Asian directional bars and North American/European truncated domes to improve navigation safety for blind and low-vision individuals.