AI/ML arXiv cs.AI

S-EMBER: A Large-Scale Benchmark for Streaming Egocentric Memory Retrieval

Introduces S-EMBER, a large-scale benchmark for streaming egocentric memory retrieval using data from Ray-Ban Meta smart glasses.

AI/ML arXiv cs.AI

LLMoxie: Exploring Agentic AI for Scientific Software Development

Describes LLMoxie, an institutional AI platform for scientific software development integrating multi-cloud inference and domain-aware coding agents.

Tech Business/VC Hacker News

Path to Google's #1 Position

A discussion regarding the factors that led to Google's dominant position in search.

Software Engineering Hacker News

Git Hash Chain Malleability

A technical exploration of malleability issues within Git hash chains.

AI/ML arXiv cs.AI

Homer: Understanding Long-form Videos with Hierarchical Memory and Agentic Reasoning

Introduction of Homer, a framework using hierarchical memory and agentic reasoning to improve LLM understanding of long-form videos.

AI/ML arXiv cs.AI

An automated method of identifying incorrectly labelled images based on the sequences of loss functions of deep learning networks

A proposed automated method to identify incorrectly labeled medical images by analyzing loss function sequences during deep learning training.

AI/ML arXiv cs.AI

AgentLTL: A Trace-Verification Framework for Measuring, Enforcing, and Training Procedural Compliance in Tool-Using LLM Agents

AgentLTL is a trace-verification framework that uses First-Order Linear Temporal Logic to enforce and train procedural compliance in tool-using LLM agents.

AI/ML arXiv cs.AI

Knowledge-Centric Information Systems

The paper proposes 'knowledge architecture' as a new discipline to manage organizational knowledge as executable infrastructure for AI systems.

AI/ML arXiv cs.AI

Fusion: A Framework for Unified Sequential Token AdaptatIon in VisiOn TraNsformers

Fusion is a unified adaptive inference framework for Vision Transformers that coordinates token merging, early exiting, and pruning to reduce energy and latency.

AI/ML arXiv cs.AI

The agent creates, we validate: A Lightweight Framework for Agentic Artifact Generation

A lightweight framework for reliable LLM artifact generation focusing on validation rigor through test-driven generation and expert-distilled judges.

AI/ML arXiv cs.AI

COMET: Combinatorial Optimization for Multiplex Editing Targets Via Constraint-Preserving QAOA

COMET compares penalty-based and XY-mixer QAOA strategies for optimizing multiplex CRISPR-Cas9 gene editing targets on quantum hardware.

Software Engineering arXiv cs.AI

Fine-Grained Computation Offload for Off-the-Shelf Servers in Tens of Lines

A method for fine-grained computation offloading on off-the-shelf servers that recovers performance by rerouting requests to avoid context switch and busy-wait costs.

Other Hacker News

Show HN: Chiptune Radio

A showcase of Chiptune Radio, a streaming service or platform dedicated to chiptune music.

Other Hacker News

Canada's only watchmaking school still ticking after 80 years

A feature story about Canada's only watchmaking school and its 80-year history.

Software Engineering Hacker News

Show HN: Free Mermaid Diagram Editor

A new free editor for creating diagrams using the Mermaid syntax.

Cybersecurity Hacker News

Tenda firmware (multiple versions) contains hidden authentication backdoor

Multiple versions of Tenda router firmware have been found to contain a hidden authentication backdoor.

AI/ML arXiv cs.AI

From Raw Segmentations to Simulation-Ready Cardiac Meshes: An Automated Framework for Anatomical Reconstruction and Virtual Cohort Generation

An automated open-source framework that converts CT segmentations into simulation-ready cardiac meshes for virtual cohort generation.

AI/ML arXiv cs.AI

CRODA-ST: Single-Target Cross-Receiver Open-Set Radio Fingerprint Recognition

Introduction of CRODA-ST, a framework for single-target cross-receiver open-set radio fingerprint recognition to improve IoT device authentication.

AI/ML arXiv cs.AI

DOSE-I: A Multimodal Biosignal Dataset of Procedural Sedation for Endoscopy -- Technical Report

A technical report on DOSE-I, a multimodal biosignal dataset of procedural sedation for endoscopy, including pEEG preprocessing C code.

AI/ML arXiv cs.AI

Additive Causal Construction for Transferable and Reconfigurable Cross-System Learning in Multi-Source Image Fusion

The ACC framework addresses cross-system discrepancy in multi-source image fusion using causal construction and uncertainty quantification for better OOD generalization.