All Articles
16590 articles total
Investigation of Polycystic Ovary Syndrome (PCOS) Diagnosis Using Machine Learning Approaches
This research investigates using various machine learning algorithms, such as AdaBoost and XGBoost, to improve the early diagnosis of Polycystic Ovary Syndrome (PCOS).
CADENCE: Closing the Reasoning Gap via Coverage-Adaptive On-Policy Distillation
CADENCE is a new distillation framework that closes the reasoning gap between large teacher and compact student LLMs using a coverage-adaptive on-policy approach.
TurboVec: A Case Study in Cost-Efficient Private Retrieval for Enterprise RAG via Codebook-Oblivious Quantization
TurboVec is an open-source vector index utilizing TurboQuant, a codebook-oblivious scalar quantizer that improves retrieval efficiency and privacy for enterprise RAG.
Real-World Evaluation of an AI Agent Drafting Translational Impact Summaries
A human-in-the-loop AI agent was developed to automate the assembly of research impact dossiers and draft summaries for clinical scholars.
Automated Cardiac Adipose Tissue Segmentation in Computed Tomography: A Literature Review
A literature review examines automated segmentation methods for cardiac adipose tissue in CT scans, highlighting the potential for clinical biomarkers.
Show HN: Workaround – Unstar GitHub Repos in Bulk
Workaround is a new tool designed to help GitHub users unstar repositories in bulk, simplifying repository management.
OpenAI says Hugging Face was breached by its pre-release models
OpenAI claimed that a breach of Hugging Face was caused by internal testing gone wrong.
Beyond Memory Leaderboards: Evaluating Scientific Memory as Budgeted Context Restoration
Researchers introduce two scientific-memory benchmarks (PAIM and PTr) and the Theoria system to evaluate LLM memory in scientific contexts.
Principled Direction-Free Intrinsic Motivation through Model-Free Epistemic Free-Energy Estimators
This paper proposes a new intrinsic reward system for unsupervised reinforcement learning to better handle exploration in environments with mixed uncertainty.
Do Speech Tokens Leak Voiceprints? Speaker Inversion Attacks Against End-to-End Speech Language Models
Researchers demonstrate that speech tokens in end-to-end speech language models can leak voiceprints, enabling speaker inversion attacks.
Trace-Based On-Policy Distillation for Masked Diffusion Language Models
Introducing TOPD, a trace-based on-policy distillation framework that efficiently transfers reasoning capabilities to masked diffusion language models.
A Deep Reinforcement Learning Algorithm for the Vehicle Routing Problem with Stochastic Demands and Outsourcing
A deep reinforcement learning approach using graph attention networks to optimize the vehicle routing problem with stochastic demands and outsourcing.
Certified-Gap Dual-Price Policies for Real-Time Truckload Bid Acceptance with Relocating, Clock-Constrained Resources
The paper presents a dual-price policy for real-time truckload bid acceptance, providing certified optimality gaps for fleet management.
A Method for Learning Value Systems in Generative AI
A method for learning human value systems in generative AI to improve alignment and explainability using multi-objective reward models.
PREFAIL: Identifying Precursors to Failures in Robotic Lift-and-Place Tasks to Improve Task Execution Performance
PREFAIL is a failure prediction method for robotic lift-and-place tasks that identifies precursors to failure to improve execution performance.
Meta is testing an AI bedtime story app for people with no imagination
Meta is testing 'StoryKit', an AI-powered app designed to help parents create bedtime stories for their children.
How Do You Choose Your AI Component? An Interview Study of Secure AI Integration in Practice
An interview study explores how software practitioners integrate AI components into systems, finding that security considerations are often ignored in favor of functional criteria.
Building a Neural Network from Scratch: Implementation, Evaluation, and Optimization
A research paper presents a self-contained neural network framework built from scratch without automatic differentiation to serve as a pedagogical tool.
OFD-Net: Teacher-Free Reliable Semi-supervised Medical Image Segmentation with Orthogonal Feature Disentanglement Net of Foreground-Background
The authors propose OFD-Net, a teacher-free semi-supervised framework for medical image segmentation using orthogonal feature disentanglement.
A Causal Markov Condition for Value
This paper introduces the value Causal Markov Condition (v-CMC) to create a mathematical foundation for causal value theory linking causality and utility.