·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
MAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation13m◆Counterfactual Model Testing Under Distribution Shift13m◆Centralized Critics for N-Agent Ad Hoc Teamwork13m◆Where and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning13m◆Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction13m◆Numina Fuse: Interactive Autoformalization in Lean 413m◆On the Limits of Sparse Linear Model with Shuffled Differential Privacy13m◆Video World Models Need Predictable Latents13m◆AutoPRL: Automated Path Representation Learning via a Learnable View Generator13m◆Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models13m◆Latent Goal Prediction from Language for Model-Based Planning13m◆Rectified Language Flow13m◆Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment13m◆Intrinsic Goals for Autonomous Agents: Hierarchical Exploration Captures Future Volitional Action in Drosophila13m◆Estimating Long-term Heterogeneous Dose-response Curve: Generalization Bound Leveraging Optimal Transport Weights13m◆Intervention-Certified Amplification: A Validity Gate for $\sigma$-Weighted Post-Training13m◆Single-Cycle Battery Remaining Useful Life Prediction with Compressed Continuous Wavelet Features and Deep Neural Networks13m◆DoubleFree:Exploiting Exploration History for Rollout-Free Multi-Task LLMs Training13m◆Unbiased Alignment for Large Language Models with Noisy Preferences13m◆ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration13m◆MAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation13m◆Counterfactual Model Testing Under Distribution Shift13m◆Centralized Critics for N-Agent Ad Hoc Teamwork13m◆Where and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning13m◆Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction13m◆Numina Fuse: Interactive Autoformalization in Lean 413m◆On the Limits of Sparse Linear Model with Shuffled Differential Privacy13m◆Video World Models Need Predictable Latents13m◆AutoPRL: Automated Path Representation Learning via a Learnable View Generator13m◆Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models13m◆Latent Goal Prediction from Language for Model-Based Planning13m◆Rectified Language Flow13m◆Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment13m◆Intrinsic Goals for Autonomous Agents: Hierarchical Exploration Captures Future Volitional Action in Drosophila13m◆Estimating Long-term Heterogeneous Dose-response Curve: Generalization Bound Leveraging Optimal Transport Weights13m◆Intervention-Certified Amplification: A Validity Gate for $\sigma$-Weighted Post-Training13m◆Single-Cycle Battery Remaining Useful Life Prediction with Compressed Continuous Wavelet Features and Deep Neural Networks13m◆DoubleFree:Exploiting Exploration History for Rollout-Free Multi-Task LLMs Training13m◆Unbiased Alignment for Large Language Models with Noisy Preferences13m◆ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration13m◆
News/A Physics-Inspired Optimizer: Velocity Regularized Adam
arxiv
PublishedJune 12, 2026 at 4:00 AM
▲bullish

A Physics-Inspired Optimizer: Velocity Regularized Adam

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2505.13196v3 Announce Type: replace Abstract: We introduce Velocity-Regularized Adam (VRAdam), a physics-inspired optimizer for training deep neural networks that draws on ideas from quartic terms for kinetic energy with its stabilizing effects on various system dynamics. Previous algorithms,

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Mentioned models
03
  • 01
    Velocity-Regularized Adam (VRAdam)
  • 02
    Adam
  • 03
    AdamW
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
03
#optimization#deep-learning#machine-learning

No replies yet. Be first.

Mentioned models
03
  • 01
    Velocity-Regularized Adam (VRAdam)
  • 02
    Adam
  • 03
    AdamW
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
03
#optimization#deep-learning#machine-learning

Related coverage

More from ARXIV
arxivMAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation13marxivCounterfactual Model Testing Under Distribution Shift13marxivCentralized Critics for N-Agent Ad Hoc Teamwork13marxivWhere and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning13m
The Bubble Brief
MON · WED · FRI

Read optimization insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews