AI Agents

[P] Easily provide Wandb logs as context to agents for analysis and planning.

It is frustrating to use the Wandb CLI and MCP tools with my agents. For one, the MCP tool basically floods the context window and freque...

Deepmind's 'AI Agent Traps' Paper Maps How Hackers Could Weaponize AI Agents Against Users

Claude, OpenClaw and the new reality: AI agents are here — and so is the chaos

[2602.17902] El Agente Gráfico: Structured Execution Graphs for Scientific Agents

The paper introduces El Agente Gráfico, a framework that enhances scientific workflows by integrating LLMs with structured execution grap...

[2602.17831] The Token Games: Evaluating Language Model Reasoning with Puzzle Duels

The Token Games introduces a novel evaluation framework for language models, using puzzle duels to assess reasoning capabilities without ...

[2602.17826] Ontology-Guided Neuro-Symbolic Inference: Grounding Language Models with Mathematical Domain Knowledge

This article explores the integration of formal domain ontologies into language models to enhance their reliability in mathematical reaso...

[2602.17676] Epistemic Traps: Rational Misalignment Driven by Model Misspecification

This paper explores how model misspecification leads to rational misalignments in AI behavior, presenting a new framework for understandi...

[2602.17931] Memory-Based Advantage Shaping for LLM-Guided Reinforcement Learning

This article presents a novel approach to reinforcement learning (RL) using memory-based advantage shaping, leveraging large language mod...

[2602.17930] MIRA: Memory-Integrated Reinforcement Learning Agent with Limited LLM Guidance

The paper presents MIRA, a Memory-Integrated Reinforcement Learning Agent that reduces reliance on large language models (LLMs) by utiliz...

[2602.17898] Breaking the Correlation Plateau: On the Optimization and Capacity Limits of Attention-Based Regressors

This paper explores the limitations of attention-based regression models, particularly the phenomenon of the Pearson correlation coeffici...

[2602.17832] MePoly: Max Entropy Polynomial Policy Optimization

MePoly introduces a novel polynomial energy-based model for policy optimization in stochastic control, enhancing multi-modality represent...

[2602.17827] Avoid What You Know: Divergent Trajectory Balance for GFlowNets

The paper presents Adaptive Complementary Exploration (ACE), an algorithm designed to enhance the efficiency of Generative Flow Networks ...

[2602.17798] Grassmannian Mixture-of-Experts: Concentration-Controlled Routing on Subspace Manifolds

The paper presents Grassmannian Mixture-of-Experts (GrMoE), a novel routing framework that enhances expert assignment in machine learning...

[2602.17744] Bayesian Optimality of In-Context Learning with Selective State Spaces

This paper introduces Bayesian optimal sequential prediction as a framework for understanding in-context learning (ICL), demonstrating it...

[2602.17695] EXACT: Explicit Attribute-Guided Decoding-Time Personalization

The paper presents EXACT, a novel approach for decoding-time personalization in large language models, enhancing user alignment through i...

[2602.17692] Agentic Unlearning: When LLM Agent Meets Machine Unlearning

The paper introduces 'agentic unlearning,' a novel approach to remove sensitive information from both model parameters and memory in AI a...

[2602.17685] Optimal Multi-Debris Mission Planning in LEO: A Deep Reinforcement Learning Approach with Co-Elliptic Transfers and Refueling

This paper presents a novel approach to multi-target active debris removal in Low Earth Orbit using deep reinforcement learning, co-ellip...

Samsung is adding Perplexity to Galaxy AI for its upcoming S26 series

Samsung is integrating Perplexity's AI agent into its Galaxy AI for the upcoming S26 series, enhancing user experience with multiple AI f...

What Happened When We Let AI Agents Cross-Examine Each Other

The article explores an experiment where AI agents cross-examined each other after a summit, revealing insights about their interactions ...

AI agent invasion leaves people scrambling to pick winners | Daily Sabah

The rise of AI agents capable of performing complex tasks is reshaping the tech landscape, prompting investors to reassess their strategi...

Samsung is adding Perplexity to Galaxy AI | The Verge

Samsung is integrating Perplexity into its Galaxy AI, enhancing its multi-agent ecosystem to allow users to interact with various AI agen...

[P] Ai Learns to play Street Fighter 6

This article details the process of training an AI to play Street Fighter 6 using imitation learning, showcasing both the gameplay and te...

We have HR for managing human capital. What's the equivalent for AI agents?

The article discusses the need for a management framework for AI agents, similar to HR for human capital, as organizations increasingly d...