AI News — Latest Developments & Breakthroughs

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

StreamGaze: Gaze-Guided Temporal Reasoning and Proactive Understanding in Streaming Videos

arXiv:2512.01707v2 Announce Type: replace-cross Abstract: Streaming video understanding requires models not only to process temporally incoming frames, but also

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

WorldMM: Dynamic Multimodal Memory Agent for Long Video Reasoning

arXiv:2512.02425v2 Announce Type: replace-cross Abstract: Recent advances in video large language models have demonstrated strong capabilities in understanding

ArXiv cs.AI 📄 Paper ⚡ AI Lesson 14h ago

Fluent Alignment with Disfluent Judges: Post-training for Lower-resource Languages

arXiv:2512.08777v2 Announce Type: replace-cross Abstract: We propose a post-training method for lower-resource languages that preserves the fluency of language

ArXiv cs.AI 👁️ Computer Vision 📄 Paper ⚡ AI Lesson 14h ago

Particulate: Feed-Forward 3D Object Articulation

arXiv:2512.11798v2 Announce Type: replace-cross Abstract: We introduce Particulate, a feed-forward model that, given a 3D mesh of an object, infers its articula

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models

arXiv:2512.13607v2 Announce Type: replace-cross Abstract: Building general-purpose reasoning models with reinforcement learning (RL) entails substantial cross-d

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

SonicMoE: Accelerating MoE with IO and Tile-aware Optimizations

arXiv:2512.14080v2 Announce Type: replace-cross Abstract: Mixture of Experts (MoE) models have emerged as the de facto architecture for scaling up language mode

ArXiv cs.AI 📄 Paper ⚡ AI Lesson 14h ago

PathFinder: Advancing Path Loss Prediction for Single-to-Multi-Transmitter Scenario

arXiv:2512.14150v3 Announce Type: replace-cross Abstract: Radio path loss prediction (RPP) is critical for optimizing 5G networks and enabling IoT, smart city,

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

Dual-objective Language Models: Training Efficiency Without Overfitting

arXiv:2512.14549v3 Announce Type: replace-cross Abstract: This paper combines autoregressive and masked-diffusion training objectives without any architectural

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

MRG-R1: Reinforcement Learning for Clinically Aligned Medical Report Generation

arXiv:2512.16145v2 Announce Type: replace-cross Abstract: Medical report generation aims to automatically produce radiology-style reports from medical images, s

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

Hearing to Translate: The Effectiveness of Speech Modality Integration into LLMs

arXiv:2512.16378v3 Announce Type: replace-cross Abstract: As Large Language Models (LLMs) expand beyond text, integrating speech as a native modality has given

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

The Dual-State Architecture for Reliable LLM Agents

arXiv:2512.20660v2 Announce Type: replace-cross Abstract: Large Language Models deployed as code generation agents exhibit stochastic behavior incompatible with

ArXiv cs.AI 👁️ Computer Vision 📄 Paper ⚡ AI Lesson 14h ago

RoAD Benchmark: How LiDAR Models Fail under Coupled Domain Shifts and Label Evolution

arXiv:2601.07855v2 Announce Type: replace-cross Abstract: For 3D perception systems to operate reliably in real-world environments, they must remain robust to e

ArXiv cs.AI 📄 Paper ⚡ AI Lesson 14h ago

Incorporating Q&A Nuggets into Retrieval-Augmented Generation

arXiv:2601.13222v2 Announce Type: replace-cross Abstract: RAGE systems integrate ideas from automatic evaluation (E) into Retrieval-augmented Generation (RAG).

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

Insider Knowledge: How Much Can RAG Systems Gain from Evaluation Secrets?

arXiv:2601.13227v2 Announce Type: replace-cross Abstract: RAG systems are increasingly evaluated and optimized using LLM judges, an approach that is rapidly bec

ArXiv cs.AI 👁️ Computer Vision 📄 Paper ⚡ AI Lesson 14h ago

CARPE: Context-Aware Image Representation Prioritization via Ensemble for Large Vision-Language Models

arXiv:2601.13622v3 Announce Type: replace-cross Abstract: Large vision-language models (LVLMs) are typically trained using autoregressive language modeling obje

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

NRR-Phi: Text-to-State Mapping for Ambiguity Preservation in LLM Inference

arXiv:2601.19933v5 Announce Type: replace-cross Abstract: Large language models exhibit a systematic tendency toward early semantic commitment: given ambiguous

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

AI and My Values: User Perceptions of LLMs' Ability to Extract, Embody, and Explain Human Values from Casual Conversations

arXiv:2601.22440v2 Announce Type: replace-cross Abstract: Does AI understand human values? While this remains an open philosophical question, we take a pragmati

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

EDU-CIRCUIT-HW: Evaluating Multimodal Large Language Models on Real-World University-Level STEM Student Handwritten Solutions

arXiv:2602.00095v2 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) hold significant promise for revolutionizing traditional educ

ArXiv cs.AI 📄 Paper 14h ago

TernaryLM: Memory-Efficient Language Modeling via Native 1.5-Bit Quantization with Adaptive Layer-wise Scaling

arXiv:2602.07374v2 Announce Type: replace-cross Abstract: Large language models (LLMs) achieve remarkable performance but demand substantial computational resou

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

PISCO: Precise Video Instance Insertion with Sparse Control

arXiv:2602.08277v2 Announce Type: replace-cross Abstract: The landscape of AI video generation is undergoing a pivotal shift: moving beyond general generation -

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

SWE Context Bench: A Benchmark for Context Learning in Coding

arXiv:2602.08316v2 Announce Type: replace-cross Abstract: Large language models are increasingly used as programming agents for repository level software engine

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

Administrative Law's Fourth Settlement: AI and the Capability-Accountability Trap

arXiv:2602.09678v2 Announce Type: replace-cross Abstract: Since 1887, administrative law has navigated a "capability-accountability trap": technological change

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

The Effective Depth Paradox: Evaluating the Relationship between Architectural Topology and Trainability in Deep CNNs

arXiv:2602.13298v2 Announce Type: replace-cross Abstract: This paper investigates the relationship between convolutional neural network (CNN) and image recognit

ArXiv cs.AI 🧠 Large Language Models 📄 Paper ⚡ AI Lesson 14h ago

DUET-VLM: Dual stage Unified Efficient Token reduction for VLM Training and Inference

arXiv:2602.18846v2 Announce Type: replace-cross Abstract: Vision-language models (VLMs) have achieved remarkable multimodal understanding and reasoning capabili

📰 AI News