Tianbao (TB) Yang @yang_ML

Professor at Texas A&M University; ML/AI researcher; optimization for ML/AI; large reasoning models, developing LibAUC library for training deep neural nets. College Station, TX Joined March 2016

Tweets

195
Followers

659
Following

1K
Likes

570

Tianbao (TB) Yang @yang_ML

2 months ago

Check this out

0 0 1 193 0

View Details

1/4 🧠⚡ Can reasoning models think faster without thinking worse? Recent systems like Meta Muse Spark use length penalties in training to reduce unnecessary output tokens. A simple approach is to combine penalty-based reward with correctness-based reward in GRPO.

2 0 1 349 0

View Details

Tianbao (TB) Yang @yang_ML

2 months ago

4/4 📊 The results are remarkable: DRPO surpasses six strong GRPO-based baselines, achieving much shorter reasoning traces while maintaining or even improving performance. 📄 Paper: arxiv.org/abs/2510.04474 💻 Code: github.com/Optimization-A…

0 0 0 79 0

View Details

Tianbao (TB) Yang @yang_ML

2 months ago

3/4 🚀 In our recent ICLR paper, we propose DRPO: Decoupled Reward Policy Optimization. 💡 Key idea: decouple length optimization for correct and incorrect rollouts, so the model learns to be concise without punishing valid reasoning.

1 0 1 103 0

View Details

Tianbao (TB) Yang @yang_ML

2 months ago

@_vztu They will hire from China

0 0 1 332 1

View Details

Tianbao (TB) Yang @yang_ML

3 months ago

@ziv_ravid That is actually icml organizer intended to do so to detect llm generated reviews.

0 0 2 775 0

View Details

Tianbao (TB) Yang @yang_ML

4 months ago

@docmilanfar So it is a local minimum

0 0 0 247 0

View Details

Tianbao (TB) Yang @yang_ML

4 months ago

@AndrewYNg People who have more friends will have high temperature parameter. The same principle applies to ML.

0 0 0 95 0

View Details

Tianbao (TB) Yang @yang_ML

4 months ago

@docmilanfar @theNAEng Congratulations!

0 0 1 62 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@ylecun @FrancoisChauba1 @agupta Texas A&M has 700+ H200 GPUs

0 0 6 2K 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@AleksandraFaust @roydanroy @iclr_conf I learned that this year ICLR did have grad students serving as AC. This information was found on the student’s website. I am not saying graduate students are not necessarily qualified for AC. If the selection process is not well done, no guarantee fair decisions solely by AC.

2 0 1 505 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@AleksandraFaust @roydanroy @iclr_conf Glad to hear that.

1 0 0 447 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@roydanroy I think they are probably nominated by senior AC

0 0 0 63 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@ericxing Thanks for this great effort. It would be highly appreciated by the community. My group will definitely explore the released datasets and code.

0 0 1 188 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

@roydanroy Yes, it actually happened for NeurIPS 2025.

2 0 1 333 0

View Details

Achleshwar Luthra @luthra_achal

6 months ago

Thank you so much for the shoutout!! @GalantiTomer If you’re excited to dig into why self-supervised contrastive learning works so well, come swing by our poster session! 📍 Exhibit Hall C/D/E — Poster #2607 🗓️ Fri, Dec 5 • 4:30–7:30 PM PST

Tomer Galanti @GalantiTomer

6 months ago

Sadly, I won't be at @NeurIPSConf this year. But my brilliant Ph.D. student @luthra_achal will be there. If you are curious what is the relationship between self-supervised and supervised learning, check out our poster!

0 0 12 2K 0

0 1 1 1K 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

2/2 Shouldn’t this be strictly prohibited from the outset? If such conflicts are not prevented, it becomes difficult to maintain trust in the review system, including designations like “oral paper” or “best paper.”

0 0 2 270 0

View Details

Tianbao (TB) Yang @yang_ML

6 months ago

1/2 Given the recent leakage of reviewer identities, some authors have reported that their papers were reviewed by people from the same institution or even the same research group. This raises serious concerns: how is such a conflict of interest even possible?