Zhehui Chen

Research@Google Deepmind;
Long horizon agentic coding: eval and data; RSI;

Ex-MLE@DiDiChuxingCo. Ph.D.@GeorgiaTechISyE.

[Google DeepMind @GoogleDeepMind]

Two new Gemini models are here to help scale your AI agents and secure code:

🔘 3.8 Flash: our most intelligent model yet with significant gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.

🔘 3.8 Flash Cyber: our most capable cybersecurity model with frontier-level vulnerability detection and automated patching.

[Tuo Zhao @tourzhao]

Hi, Reviewer,

If you have such a concern with the review system of NeurIPS, I would encourage you to contact the program chairs.

Why do you want to penalize the authors by reducing the review rating to 5?

BTW, I've provided a proof sketch in the paper. 😂

Best,

Authors

[Tuo Zhao @tourzhao]

Can GANS accurately recover data distributions? How large should the generator networks be for sufficiently expressing distributions?

Check our paper - https://t.co/DP9kfCQHjj, where we establish the statistical rates of convergence for distribution estimation using GANs.

[Tuo Zhao @tourzhao]

Haoming's (@jiang_haoming ) presentation of this work on ACL 2020 can be found at

https://t.co/aRwIdCyutO

[Tuo Zhao @tourzhao]

Check our ACL 2020 paper on Adaptive Knowledge Sharing across Multiple Domains in Neural Machine Translation: https://t.co/pNDC0HYDBL

The key is the word-level and layer-wise domain mixing layer!

Joint work by @GeorgiaTechISyE and @BytedanceTalk

@jiang_haoming @Ashleylololol

[Tuo Zhao @tourzhao]

Ever been curious about why ResNets are easier to train than Feedforward Networks?

Check our Paper "Towards Understanding the Importance of Shortcut Connections in Residual Networks".

https://t.co/fEpQ2BnRbK

@SimonShaoleiDu

[Tuo Zhao @tourzhao]

Can noise really help escape from bad optima and converge to a good one during the training of deep neural networks?

Check our paper "Towards Understanding the Importance of Noise in Training Neural networks" -- https://t.co/kd6S8S1CmO

[Tuo Zhao @tourzhao]

Struggling with fine-tuning BERT models? Overfit your tasks again?

Check our recent work on ACL 2020 "SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization" https://t.co/JO2pxb8SiD

[Tuo Zhao @tourzhao]

Frustrated by your LSTM-based point process models? Fail to capture long-term dependencies? Cannot scale to large data well?

Check our recent paper on ICML 2020 -- “Transformer Hawkes Process” at https://t.co/X5kpOfVTpj

@SimiaoZuo @jiang_haoming