cd/entity/Zhu et al.· home› entities› Zhu et al.
grep -l @zhu et al. /news/*.json | wc -l → 3

Zhu et al.

mentions 3 type Person feed RSS

// recent coverage 3 mentions

00:00
2026-09-11
g-ftech.com
machine-learning

We Tried ISO-AdamW. AdamW Kept Its Job.

A controlled pilot on a dedicated NVIDIA H200 GPU found ISO-AdamW scored 758 correct answers versus 754 for baseline AdamW on a 1,000-question held-out test, a +0.4% delta with a 95% confidence interv…

00:00
2026-09-08
g-ftech.com
machine-learning

SFT vs. RL: What Changes Inside the Model?

New research by Zhu et al. (July 2026, arXiv:2607.19331, "ISO: An RLVR-Native Optimization Stack") provides mathematical proof that Reinforcement Learning with Verifiable Rewards (RLVR) post-training …

07:42
2026-05-26
djdumpling.github.io
machine-learning

paper reading catalog

DeepSeek researchers introduced manifold-constrained hyper-connections to restore the identity mapping property in transformer architectures, addressing training instability and scalability issues cau…

// co-occurs with top 8 entities
GRPO 2 ISO: An RLVR-Native Optimization Stack 2 DeepSeek 1 ISO-AdamW 1 AdamW 1 NVIDIA H200 1 GSM8K 1 Qwen 1
// topics top 5 topics
machine learning 3 ai research 3 large language models 2 neural networks 1 artificial intelligence 1