This paper explores how reinforcement learning agents can use environmental features, termed artifacts, to function as external memory. By formalizing this intuition within a mathematical framework, the authors prove that certain observations can reduce the information required to represent an agent's history. Through experiments with spatial navigation tasks using both Linear Q-learning and Deep Q-Networks (DQN), the study demonstrates that observing paths or landmarks allows agents to achieve higher performance with lower internal computational capacity. Notably, this effect of externalized memory emerges unintentionally through the agent's sensory stream without explicit design for memory usage.
- Formalization of artifacts as observations that encode information about the past.
- The Artifact Reduction Theorem proving environmental artifacts reduce history representation requirements.
- Empirical evidence showing reduced internal capacity needs when spatial paths are visible.
- Observation that externalized memory can emerge implicitly in standard RL agents.
- Implications for agent design, suggesting performance gains may come from environment-agent coevolution rather than just scaling parameters.
- Formalization of artifacts as observations that encode information about the past.
- The Artifact Reduction Theorem proving environmental artifacts reduce history representation requirements.
- Empirical evidence showing reduced internal capacity needs when spatial paths are visible.
- Observation that externalized memory can emerge implicitly in standard RL agents.
- Implications for agent design, suggesting performance gains may come from environment-agent coevolution rather than just scaling parameters.
AutoAgent is a revolutionary open-source library designed to automate the tedious process of agent engineering and prompt tuning. By employing a meta-agent, the library allows for the autonomous optimization of an agent's harness, including system prompts, tool definitions, and orchestration strategies, all without human intervention. During a 24-hour run, AutoAgent achieved impressive results, including the top score on SpreadsheetBench and a leading GPT-5 score on TerminalBench. This technology effectively transitions the human's role from a manual engineer to a high-level director, enabling rapid, self-improving agent development across various domains and benchmarks.
"The article discusses the evolution of manufacturing beyond 'smart' to an AI-driven future. It argues that while smart manufacturing focused on connectivity and data collection, AI will unlock true transformation by enabling predictive maintenance, optimized supply chains, and personalized product development. The piece outlines ten specific use cases where AI is poised to make a significant impact, including generative design, digital twins, and autonomous quality control. It emphasizes the shift from reactive problem-solving to proactive optimization, ultimately leading to increased efficiency, reduced costs, and improved product quality. The author posits that AI is not just enhancing manufacturing, but fundamentally reshaping it."
pi-autoresearch is an autonomous experiment loop for optimizing various targets like test speed, bundle size, LLM training, or build times. Inspired by karpathy/autoresearch, it utilizes a skill-extension architecture, allowing domain-agnostic infrastructure paired with domain-specific knowledge. The core workflow involves editing code, committing changes, running experiments, logging results, and either keeping or reverting the changes – a cycle that repeats indefinitely. Key components include a status widget, a detailed dashboard, and configuration options for customizing behavior. It persists experiment data in `autoresearch.jsonl` and session context in `autoresearch.md` for resilience and reproducibility.
This project, `autoresearch-opencode`, is an autonomous experiment loop designed for use with OpenCode. It's a port of `pi-autoresearch`, but implemented as a pure skill, eliminating the need for an MCP server and relying solely on instructions the agent follows using its built-in tools. The skill allows users to automate optimization tasks, as demonstrated by the example of optimizing the BogoSort algorithm which achieved a 7,802x speedup by leveraging Python's `bisect` module for sorted-state detection.
The system maintains state using a JSONL file, enabling resume/pause functionality and detailed experiment tracking. It provides a dashboard for monitoring progress and ensures data integrity through atomic writes and validation checks.
The system maintains state using a JSONL file, enabling resume/pause functionality and detailed experiment tracking. It provides a dashboard for monitoring progress and ensures data integrity through atomic writes and validation checks.
This article explores five Python decorators that can be used to optimize LLM-based applications. These decorators leverage libraries like functools, diskcache, tenacity, ratelimit, and magnetic to address common challenges such as caching, network resilience, rate limiting, and structured output binding. The article provides code examples to illustrate how each decorator can be implemented and used to improve the performance and reliability of LLM applications.
MIT researchers developed a new approach that rethinks how a classic method, known as Bayesian optimization, can be used to solve problems with hundreds of variables. In tests on realistic engineering-style benchmarks, like power-system optimization, the approach found top solutions 10 to 100 times faster than widely used methods.
Their technique leverages a foundation model trained on tabular data that automatically identifies the variables that matter most for improving performance, repeating the process to hone in on better and better solutions. The researchers’ tabular foundation model does not need to be constantly retrained as it works toward a solution, increasing the efficiency of the optimization process.
The technique also delivers greater speedups for more complicated problems, so it could be especially useful in demanding applications like materials development or drug discovery. The research will be presented at the International Conference on Learning Representations.
Their technique leverages a foundation model trained on tabular data that automatically identifies the variables that matter most for improving performance, repeating the process to hone in on better and better solutions. The researchers’ tabular foundation model does not need to be constantly retrained as it works toward a solution, increasing the efficiency of the optimization process.
The technique also delivers greater speedups for more complicated problems, so it could be especially useful in demanding applications like materials development or drug discovery. The research will be presented at the International Conference on Learning Representations.
This research introduces Doc-to-LoRA (D2L), a method for efficiently processing long documents with Large Language Models (LLMs). D2L creates small, adaptable "LoRA" modules that distill key information from a document, allowing the LLM to answer questions without needing the entire document in memory. This significantly reduces latency and memory usage, enabling LLMs to handle contexts much longer than their original capacity and facilitating faster knowledge updates.
Researchers have refined the simplex method, a key algorithm for optimization, proving it can't be improved further and providing theoretical reasons for its efficiency.
CUDA Tile is a new Python package that simplifies GPU programming by automatically tiling loops, handling data transfer, and optimizing memory access. It allows developers to write concise and readable code that leverages the full power of NVIDIA GPUs without needing to manually manage the complexities of parallel programming.
