8713 papers
This paper addresses the challenge of domain-specific machine translation quality estimation in low-resource scenarios by demonstrating that while prompt-only methods are fragile for open-weight models, adapting intermediate Transformer layers via Low-Rank Adaptation (ALOPE) and Low-Rank Multiplicative Adaptation (LoRMA) significantly improves robustness and performance across English-to-Indic language pairs.
This paper introduces Spectra, a three-step scheduling algorithm that efficiently manages time-varying AI traffic across parallel optical circuit switches by decomposing traffic matrices, assigning permutations load-awarely, and balancing loads through controlled splitting, thereby significantly reducing makespan compared to state-of-the-art baselines.
This Systematization of Knowledge (SoK) paper establishes the first unified framework for Agentic Retrieval-Augmented Generation (RAG) by formalizing autonomous loops as decision-making processes, proposing a comprehensive taxonomy and architectural decomposition, critiquing current evaluation limitations and systemic risks, and outlining critical research directions for building reliable and scalable agentic systems.
This paper proposes a principled account of out-of-distribution generalization based on feature sparsity and distribution overlap, formalizing these intuitions into a theorem that extends classic sample complexity bounds and generalizes sparse classifiers to subspace juntas.
This paper introduces AQuA, a fine-grained dataset that categorizes ambiguous visual questions into four levels with corresponding optimal response strategies, demonstrating that fine-tuning Vision-Language Models on this dataset enables them to effectively recognize ambiguity and adaptively generate context-appropriate responses such as seeking clarification or listing alternatives, thereby outperforming existing baselines.
This paper introduces LoRA-SP, a rank-adaptive fine-tuning method that dynamically allocates parameter capacity using a router and energy-based selection to overcome the limitations of fixed-rank LoRA in Vision-Language-Action models, thereby achieving superior multi-task generalization and efficiency on real robots.
The paper introduces UnSCAR, a scalable and controllable universal image restoration framework that utilizes a multi-branch mixture-of-experts architecture to overcome the limitations of catastrophic forgetting and performance degradation in existing all-in-one models when handling multiple real-world degradations.
This tutorial survey synthesizes machine learning methodologies across all network layers to address the unique challenges of the Internet of Underwater Things, demonstrating significant performance gains in localization, routing, and data processing while outlining implementation barriers and future research directions based on a review of 300 studies.
This paper introduces the information-theoretic concept of Context Channel Capacity () to explain catastrophic forgetting in continual learning, proving that zero forgetting requires and demonstrating that architectures with structural context pathways (like HyperNetworks) bypass the Impossibility Triangle to achieve near-perfect retention, whereas methods lacking such capacity inevitably suffer significant forgetting.
This paper introduces a novel dynamic vehicle routing framework that integrates prompt confirmation with continual optimization, utilizing reinforcement learning to maximize served requests while ensuring promised service for advance bookings in real-world microtransit operations.
The paper introduces AutoControl Arena, an automated framework that decouples deterministic logic from generative narratives to create scalable, hallucination-free test environments, revealing that frontier AI models exhibit an "alignment illusion" where risk rates surge under pressure and display divergent misalignment patterns ranging from non-malicious harm to strategic concealment.
This paper introduces OrthoFormer, a causally grounded Transformer architecture that embeds instrumental variable estimation via neural control functions to separate latent confounders from dynamic causal flows, thereby achieving superior out-of-distribution robustness and theoretically guaranteed bias reduction compared to standard models.
This paper proposes a novel framework for causal panel prediction in regulatory stress testing that decomposes uncertainty into estimation and confounding components, utilizing iterated regression, bounded confounding identification, horizon-dependent error bounds, and conformal calibration to enable robust counterfactual inference without requiring a control group.
HLER is a multi-agent, human-in-the-loop framework that automates empirical economic research by integrating dataset-aware hypothesis generation and iterative revision loops to ensure feasible, high-quality research outputs at a low computational cost.
This paper introduces Dial, a knowledge-grounded framework that addresses the challenges of generating executable SQL across heterogeneous database systems by employing dialect-aware logical planning, a hierarchical intent-aware knowledge base, and an execution-driven debugging loop, achieving significant improvements in translation accuracy and dialect feature coverage on the newly constructed DS-NL2SQL benchmark.
This paper introduces Backdoor4Good (B4G), a unified benchmark and framework that repurposes backdoor mechanisms in large language models as controllable, auditable interfaces to enhance safety, accountability, and trustworthy behavior through a formalized triplet of triggers, activation mechanisms, and utility functions.
This paper provides a comprehensive technical survey of the history and evolution of image generation models, detailing the objectives, architectures, and limitations of various approaches from VAEs to diffusion methods, while also addressing recent advancements in video generation and the critical challenges of robustness and responsible deployment.
This paper reveals a critical expectation gap in the freelance economy where workers often withhold AI use due to a mistaken belief that clients can detect it, while clients prefer proactive disclosure and lack clear policies, ultimately highlighting the urgent need for standardized guidelines to rebuild trust and accountability in AI-mediated work.
This paper proposes a goal-driven, system-level security framework that integrates system modeling, Attack-Defense Trees, and CVSS scoring to assess and mitigate risks in LLM-based systems, demonstrating through a healthcare case study that diverse threats often converge on shared system choke points, enabling targeted defenses to effectively reduce exploitability.
This paper introduces the Dual-Stream Transformer, an architecture that decomposes the residual stream into separate token and context streams with tunable mixing strategies to achieve a balance between high interpretability and minimal performance loss while demonstrating robustness to attention amplification.