Oct 25, 2026
A codebook-free image signature revisits Fisher vectors for scene retrieval
Multi-scale CNN features are aggregated into a compact second-order representation and tested on Holidays and UKBench.
Explore Franklin AI
AI Research Papers
Understand new AI research with readable explanations and links to the original papers.
Oct 25, 2026
RIPPLE trains coding agents to turn a critique into a corrective action
Recovery-focused training separates reflection, diagnosis and correction, with gains tied to available evidence and reversibility.
Oct 25, 2026
A decision-theoretic framework asks what an agent’s memory is worth
The study separates retained information, usable memory and the cost of recovering or revising evidence.
Oct 25, 2026
T2Quant allocates diffusion-model precision by task and denoising step
A label-free sensitivity signal guides mixed-precision activation quantization under a fixed bit-width budget.
Oct 25, 2026
Post-Audit Mirage studies what an offline verifier cannot identify
Matched agent lifecycles expose the same audit evidence while producing opposite deployment outcomes.
Oct 25, 2026
SetCommit coordinates parallel token choices in diffusion language models
A training-free decoder weighs token reliability against the risk of committing an inconsistent group too early.
Oct 25, 2026
PointerBench tests the precise GUI targets that broad task scores can hide
A diagnostic benchmark covers spreadsheet sub-elements, text characters and interfaces from 100 professional applications.
Oct 25, 2026
Reward-weighted preferences train solvers to value larger improvements
RWPO preserves reward-gap information when learning which elements to select in combinatorial optimization problems.
Oct 25, 2026
A tree-anchored forecast ensemble reduces photovoltaic prediction error
Selective weighting improves a strong tree baseline, but the study reuses an inspected evaluation period.
Oct 25, 2026
Trion-1 puts protein sequence, structure and language in one diffusion model
A multimodal protein model studies folding, design and cross-modal understanding in a shared token space.
Good to know
How do I use these AI research explainers?
Start with the overview to understand the paper’s main idea, then follow its original paper link for the methods, results and limitations.
Can I suggest a research paper?
Use Submit Research to send the original paper URL and explain why it is useful. An editor reviews the suggestion before publication.