Field notes
1,406 notes from my AI research vault, mapped by topic. Pick a topic to see its notes, or search.
Click the web, then scroll to zoom
Every note is in the list below. The web of topics draws here once the page has loaded.
1,406 notes and 240 topics.
- Question answered
- Source note
- Topic
- Shared notes
All notes, newest first
1,406 notes, page 45 of 59
What LLM Agents Say When No One Is Watching: Social Structure and Latent Objective Emergence in Multi-Agent Debates
Researchers studied how language models (LLMs) behave in socially structured settings, where role, audience, and relationships can influence what the LLM says publicly versus privately. They found that when there's an…
What Types of Human-AI Teams Exist?
The authors analyze 53 papers about human-AI teams and categorize them into five main clusters based on psychological taxonomies of teaming.
World Wide Models: Literary Tools for Cultural AI
The paper proposes a framework for developing culturally literate AI by integrating literary tools such as comparative reading, narratology, critical theory, world literature, and translation. It emphasizes the…
WorldSample: Closed-loop Real-robot RL with World Modelling
The paper presents WorldSample, a framework that uses reinforcement learning (RL) to improve the performance of robots by creating synthetic transitions based on real rollouts and a post-trained world model. This…
Adversarial Pragmatics for AI Safety Evaluation: A Benchmark for Instruction Conflict, Embedded Commands, and Policy Ambiguity
The text introduces 'adversarial pragmatics', a new benchmark and annotation protocol to evaluate the behavior of language models in various ambiguous situations. The goal is to clarify whether failures occur due to…
Agentic generation of verifiable rules for deterministic, self-expanding reaction classification
Researchers have developed an automated pipeline using large language models (LLMs) that can classify reactions in chemistry and generate rules for these reactions, expanding a standard taxonomy from 68 to 14,073…
Are Performance-Optimization Benchmarks Reliably Measuring Coding Agents?
The paper evaluates three performance-optimization benchmarks (GSO, SWE-Perf, and SWE-efficiency) used to assess coding agents. The authors find that these benchmarks have issues with stability, scoring rules, and the…
AutoMem: Automated Learning of Memory as a Cognitive Skill
Researchers developed AutoMem, a framework for improving a language model's memory management skills in long-horizon games. By treating memory as a trainable skill, they allowed the model to decide how to manage its…
Autonomous Scientific Discovery via Iterative Meta-Reflection
Researchers present DiscoPER, a framework for autonomous scientific discovery using large language models. It can explore datasets without predefined research objectives, conduct statistical testing to ensure validity…
Bayesian Uncertainty Propagation for Agentic RAG Pipelines: A Proof-of-Concept Study on Multi-Hop Question Answering
This research introduces an uncertainty-aware framework for Agentic Retrieval-Augmented Generation (RAG) systems, which can help identify potential failures in multi-stage reasoning pipelines. The study evaluates this…
Behavior-Adaptive Conversational Agents: Toward a Fluid Personality Framework
The authors propose a Fluid Personality Framework for conversational agents to adapt their metaphorical personas and personality expression levels based on the task context, user goals, and situational urgency.
Can Agents Generalize to the Open World? Unveiling the Fragility of Static Training in Tool Use
Researchers introduce OpenAgent, a problem setting to test how well AI agents can generalize their skills from static benchmarks to real-world scenarios with dynamic user queries and tools. The study finds that both…
CausalMix: Data Mixture as Causal Inference for Language Model Training
The paper presents CausalMix, a method that addresses the limitation of existing data mixing methods in Large Language Model (LLM) training by casting it as a causal inference problem. By using causal modeling to…
Cheap Code, Costly Judgment: A Case Study on Governable Agentic Software Engineering
In the study titled 'Cheap Code, Costly Judgment', researchers examined a case study of a 12-week development effort where an expert software engineer used AI coding agents to build a document accessibility remediation…
DART-VLN: Test-Time Memory Decay and Anti-Loop Regularization for Discrete Vision-Language Navigation
The researchers present DART-VLN, a test-time control framework for vision-language navigation agents that aims to improve their performance by addressing two common issues: stale historical evidence at memory readout…
Detecting the Undetectable: Enhancing Unsupervised time series Anomaly Detection via Active Learning
The authors propose a novel framework using active learning to enhance unsupervised time series anomaly detection in complex data. The framework aims to improve the ability of models to detect subtle and noisy…
Distill to Detect: Exposing Stealth Biases in LLMs through Cartridge Distillation
The paper introduces Distill to Detect (D2D), a method that helps identify hidden biases in language models used in high-stakes roles by amplifying these biases through cartridge distillation. The method is designed to…
Exploring the Semantic Gap in Agentic Data Systems: A Formative Study of Operationalization Failures in Analytical Workflows
A study identifies common failures in generating and executing analytical workflows using large language models across finance, human resources, and public safety domains. The authors found five recurring classes of…
From Personas to Plot: Character-Grounded Multi-Agent Story Generation for Long-Form Narratives
Researchers have developed MAGNET, a framework that generates consistent and coherent long-form narratives by using a unified approach of multi-agent goal-driven storytelling and verification. The system was tested…
FurnitureVLA: Learning Long-Horizon Bimanual Furniture Assembly with Vision-Language-Action Model
The paper introduces FurnitureVLA, a model that aims to improve long-horizon bimanual furniture assembly using Vision-Language-Action models (VLAs) in real-scale settings. The model is designed to predict actions and a…
GPU-Parallel Linearization Error Bounds for Real-Time Robust Optimal Control of Nonlinear and Neural Network Dynamics
The paper presents a new method called GPUSLS-LEO for real-time robust optimal control of uncertain nonlinear systems, specifically focusing on systems with linear time-varying (LTV) approximations. This method…
Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination
Researchers developed Graph-PRefLexOR, a model that uses graph-native reinforcement learning to generate scientifically valid hypotheses through multi-step reasoning in materials science and mechanics. The new model…
Human-Machine Collaboration on Generative Meta-Learning: Model and Algorithm
Researchers have proposed a new method called Generative Meta-Learning with Human Feedback (GMHF) to improve the performance of machine learning models in environments different from their training distribution…
Language-Critique Imitation Learning from Suboptimal Demonstrations
A new approach called Language-Critique is proposed for imitation learning from suboptimal demonstrations. Instead of using compressed signals, it utilizes natural language to provide explicit feedback about task…
An AI research agent drafts these notes from public papers and posts, and each one keeps its source link. Only the folders I choose to publish leave my Obsidian vault.
