Signals 4 · free daily AI digest

AdamX: Cosine similarity meets gradient descent

Paper recorded by Signals 4 on 2026-09-10 in cs.LG. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-10 on arXiv · recorded by Signals 4 on 2026-09-11

Category: cs.LG · 机器学习 · first seen 2026-09-11

Abstract

We introduce AdamX, a first-order optimizer that incorporates cosine similarity as an adaptive mechanism for controlling update magnitudes. The proposed method is scalable, model-agnostic, and straightforward to integrate into existing training pipelines. We further introduce a variance rectification scheme that promotes smoother optimization during the early stages of training. Overall, we provid

Read on arXiv →

#74 most recent of 215 cs.LG papers we have recorded · ↑ newer: Evaluating Time-Series Foundation Models and Multimodal Dietary Contex · ↓ older: Near-Optimal Reinforcement Learning with Multi-Step Transition Lookahe
Cite this page: AdamX: Cosine similarity meets gradient descent: the #74 most recent of 215 cs.LG papers we have recorded (as of 2026-09-10). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/adamx-cosine-similarity-meets-gradient-descent.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.LG papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions