Explore Franklin AI
Understand new AI research with readable explanations and links to the original papers.
PivotOPD targets decisive mistakes and the states they create. Its authors report improved multi-turn benchmark performance by combining preventive training with examples of how to
A controlled study finds that a minimal coding-agent session can match elaborate autonomous ML engineering harnesses under matched model, hardware and time budgets. Its cost analys
ScholarEvolve organizes published methods into candidate improvements for agent harness modules. Its reported benchmark gains test the current search procedure; continual improveme
A reanalysis finds slower learning rather than no learning, and shows that sampling budgets and aggregation rules can change which prompts receive a difficulty label.
Turbo Harness reuses earlier optimization traces to train a small editor that tailors an agent harness per instance. The paper reports improvements across seven benchmarks under sp
The path-finding framework shares local observations across a team and estimates nearby blockage risks. Its benchmark gains depend on a fixed environment, noiseless sensing and a s
A neural-operator preprint studies whether noisy Monte Carlo simulations can train useful transport surrogates, with sharply different speed and cost results across neutron and lig
Pixels to Keys: Exploring Spatial and Motion Cues in Gameplay Inverse Dynamics investigates whether an AI model can infer a player’s keyboard inputs from gam...
You Cannot Pick a Provider From the Price List: Market-Aware Routing for Open-Weight LLM Inference Choosing an LLM for a request is only part of the deployme...
Video-RSI: Recursive Self-Improvement of Video Understanding Agents via Harness Evolution Video-RSI explores whether a video-understanding agent can improve...
Start with the overview to understand the paper’s main idea, then follow its original paper link for the methods, results and limitations.
Use Submit Research to send the original paper URL and explain why it is useful. An editor reviews the suggestion before publication.
