Link: github.com/openai/math P.S.1: This reminds me of the Chinese room argument, though only indirectly. P.S.2: I imagine a few people at OpenAI have looked at the results. Also maybe some other people have looked at some papers since ~1h ago. These are o(eps) terms in the wh…
GitHub - openai/math github.com
AI Weekly's analysis
→
- OpenAI released 722 manuscripts organized into 372 families, produced by an unreleased internal model given roughly 4,000 problems.
- Each result used, on average, three hours of ChatGPT Pro thinking compute with that model.
- Not all manuscripts carry Lean formalizations, and OpenAI warns some unformalized results 'could have issues.'
Read full analysis →
View on Bluesky ·
♥ 3
↻ 0
↩ 0
·
11 from the directory shared this ·
15d ago
Temporal Difference Learning for Diffusion Models (ICML 2026) arxiv.org/abs/2606.15048 By Yangchen Pan (my former PhD student) and co-authors. It reformulates diffusion training as a Markov reward process and introduces a TD objective to encourage temporal consistency across d…
Temporal Difference Learning for Diffusion Models arxiv.org
AI Weekly's analysis
→
- The paper introduces a temporal difference objective that penalises inconsistency across the full denoising trajectory rather than only at adjacent time steps.
- It reframes diffusion as a Markov reward process and denoising as a policy evaluation problem, unifying discrete-time and continuous-time formulations.
- Reported FID gains are strongest when the number of sampling steps is small, the regime where few-step samplers and low-compute serving live.
Read full analysis →
View on Bluesky ·
♥ 14
↻ 2
↩ 1
·
2 from the directory shared this ·
119d ago
↻
Amir-massoud Farahmand reposted
@continual-rl.bsky.social
📚 Interested in Continual RL but not sure where to start? 👇 Dive in: sites.google.com/view/continu... ♾️ We've curated a resource hub with key papers, benchmarks, codebases, tutorials, and more to help you get up to speed quickly. #ContinualRL #ReinforcementLearning #MachineLe…
Continual RL Workshop - Resources sites.google.com
View on Bluesky →
There are a few more recent work, but as far as I recall, they are still under restrictive assumptions. There is a new paper came last month, but I haven't read it yet. Convergence of Monte Carlo Optimistic Policy Iteration: Beyond Uniform State-Action Updates arxiv.org/abs/26…
Convergence of Monte Carlo Optimistic Policy Iteration: Beyond Uniform State-Action Updates arxiv.org
There is John Tsitsiklis's work from 2002, but that's for the initial state update model, which feels sample inefficient. Moreover, it assumes either synchronous updates (all states are updated at the same time) or asynchronous but random with uniform distribution. www.jmlr.or…
jmlr.org
↻
Amir-massoud Farahmand reposted
Claas Voelcker
@cvoelcker.bsky.social
⏰ The deadline for our workshop workshop.oopsie-data.com on imperfect data at CoRL is nearing! ⏰ While you 🤖 are making progress on submissions, we came up with cool awards 🏆. Our workshop on imperfect data will award not only papers but also data submissions. The categories y…
Oops, I Erred — Collecting, Curating and Using Imperfect Data for Realistic Scenarios workshop.oopsie-data.com
View on Bluesky →
