Research Feed Page 27

The paper introduces a decision-aware ranking method for supply chains that selects interventions by balancing recovered net value against operational costs.

The paper analyzes six years of TrustNLP workshop proceedings to map how research has shifted from post-hoc model interpretability to proactive control of generative systems.

The paper provides a selection rule to optimally allocate a fixed memory budget between batch size and negative samples when training large scale recommender systems.

MaskFlow provides a method for accurate and seamless regional image editing by incorporating user masks directly into the generation process and using a gradient domain refinement module.

VoxSumm introduces a new corpus and framework for simultaneously summarizing and translating long-form spoken news content.

StreamFlow reduces video redundancy and improves memory efficiency by using a dynamic caching system that selectively retrieves relevant visual data during inference.

The REAP system improves closed-book knowledge base construction by using a two-stage process of relation-aware prompting and hybrid parsing to extract structured data from large language models.

ReLTEx improves automated taxonomy expansion by using LLMs for candidate generation combined with a structure-aware classifier to ensure hierarchical consistency.

The paper introduces a flexible backdoor paradigm that enables dynamic, post-training control over a Vision Language Model output by injecting trigger patterns into training data.

ConRub-Med enhances medical question answering by using automated consensus rubrics to improve reinforcement learning feedback for model responses.

The paper introduces MUSE, a large-scale knowledge base of 36,960 structured problem-solution-rationale triplets extracted directly from full-text scientific papers.

CapProbe is a benchmark for assessing detailed image captions by using region-aligned factual questions to verify specific visual content.

The paper introduces a confidence-aware training framework that aligns medical diagnostic predictions with actual accuracy to reduce clinical decision errors.

The paper introduces a relative 4D scene graph memory system to help AI assistants answer object-centric questions in long egocentric videos.

The paper introduces a new image compression architecture that disperses information across packets to maintain stable visual quality even when network connections drop data.

ImpactHO improves LLM performance during user handovers by intelligently prioritizing and transferring essential parts of the KV cache over constrained network links.

AlbumentationsX provides a centralized framework to apply synchronized image transformations across images and associated annotations.

The paper introduces Structural Logic Tensor Networks (sLTN) to enable logic-based reasoning over sequential or connected data structures by treating positional axes as primary components.

The paper introduces a corpus-free framework and benchmark for deleting specific person-related knowledge from multimodal large language models without needing the original training data.

The paper investigates whether latent activations from sparse autoencoders function as meaningful, additive components for representing conceptual similarity.