AI Agents

NeuBird AI Raises $19.3 Million To Scale Agentic AI

I got tired of 3 AM PagerDuty alerts, so I built an AI agent to fix cloud outages while I sleep. (Built with GLM-5.1)

If you've ever been on-call, you know the nightmare. It’s 3:15 AM. You get pinged because heavily-loaded database nodes in us-east-1 are ...

CodeGraphContext - An MCP server that converts your codebase into a graph database

CodeGraphContext- the go to solution for graph-code indexing 🎉🎉... It's an MCP server that understands a codebase as a graph, not chunks ...

[2511.17673] Bridging Symbolic Control and Neural Reasoning in LLM Agents: Structured Cognitive Loop with a Governance Layer

This article introduces the Structured Cognitive Loop (SCL) architecture for large language model (LLM) agents, addressing key architectu...

[2510.19771] Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents

The paper presents PROBE, a new framework for measuring proactive problem-solving capabilities in LLM agents, highlighting their limitati...

[2508.12026] Bongard-RWR+: Real-World Representations of Fine-Grained Concepts in Bongard Problems

The paper presents Bongard-RWR+, a dataset designed to enhance fine-grained visual reasoning in Bongard Problems using real-world images ...

[2510.00167] Drones that Think on their Feet: Sudden Landing Decisions with Embodied AI

The paper discusses how embodied AI enables drones to make adaptive landing decisions in real-time, enhancing their resilience and safety...

[2505.16928] Beyond Needle(s) in the Embodied Haystack: Environment, Architecture, and Training Considerations for Long Context Reasoning

This article presents the $ ext{∞-THOR}$ framework for long-horizon embodied tasks, focusing on enhancing long-context reasoning in AI th...

[2505.08021] The Correspondence Between Bounded Graph Neural Networks and Fragments of First-Order Logic

This paper explores the relationship between Bounded Graph Neural Networks (GNNs) and fragments of first-order logic, providing insights ...

[2503.17338] Capturing Individual Human Preferences with Reward Features

The paper discusses a new approach to modeling individual human preferences in reinforcement learning, emphasizing the need for adaptive ...

[2502.13062] AI-Assisted Decision Making with Human Learning

This paper explores AI-assisted decision-making, focusing on how algorithms can enhance human learning through feature selection, balanci...

[2410.13957] Goal Inference from Open-Ended Dialog

The paper discusses a method for embodied AI agents to infer user goals from open-ended dialogues using Large Language Models (LLMs), emp...

[2412.18899] GAI: Generative Agents for Innovation

The paper explores GAI, a framework for generative agents that enhances collective reasoning to foster innovation, evaluated through a ca...

[2602.17658] MARS: Margin-Aware Reward-Modeling with Self-Refinement

The paper presents MARS, a novel margin-aware reward modeling framework that enhances training efficiency by focusing on ambiguous prefer...

[2602.17641] FAMOSE: A ReAct Approach to Automated Feature Discovery

The paper presents FAMOSE, a novel framework that utilizes the ReAct paradigm for automated feature discovery in machine learning, enhanc...

[2602.17632] SMAC: Score-Matched Actor-Critics for Robust Offline-to-Online Transfer

The paper presents SMAC, a novel offline reinforcement learning method that enhances the transition from offline to online learning witho...

[2602.17616] Stable Asynchrony: Variance-Controlled Off-Policy RL for LLMs

The paper presents VCPO, a method to stabilize off-policy reinforcement learning for large language models, addressing high variance issu...

[2602.17605] Adapting Actively on the Fly: Relevance-Guided Online Meta-Learning with Latent Concepts for Geospatial Discovery

This paper presents a novel framework for geospatial discovery that integrates active learning and online meta-learning, focusing on rele...

[2602.17550] MASPO: Unifying Gradient Utilization, Probability Mass, and Signal Reliability for Robust and Sample-Efficient LLM Reasoning

The paper presents MASPO, a novel framework that addresses inefficiencies in existing Reinforcement Learning with Verifiable Rewards (RLV...

[2602.17526] The Anxiety of Influence: Bloom Filters in Transformer Attention Heads

This article explores how certain transformer attention heads act as membership testers, identifying token repetition across various lang...

[2602.17410] Improving LLM-based Recommendation with Self-Hard Negatives from Intermediate Layers

This paper presents ILRec, a novel framework that enhances LLM-based recommendation systems by utilizing self-hard negative signals from ...

[2602.17395] SpectralGCD: Spectral Concept Selection and Cross-modal Representation Learning for Generalized Category Discovery

The paper presents SpectralGCD, a novel approach for Generalized Category Discovery (GCD) that enhances multimodal learning by efficientl...

[2602.17394] Voice-Driven Semantic Perception for UAV-Assisted Emergency Networks

The paper presents SIREN, an AI framework for enhancing UAV-assisted emergency networks by converting voice communications into structure...