@askalphaxiv - AI Dynamics

Presentation of arXiv paper replications by auto-research agents

By

–

17 June 2026 17h40

Introducing paper replications for arXiv

Research codebases are notoriously hard to set up

We deploy autoresearch agents to ingest popular arXiv repos, resolve setup issues, and get the core claim running. You can now sort papers by ease of implementation 🚀 pic.twitter.com/Zw6DkX4ycq
— alphaXiv (@askalphaxiv) 17 juin 2026

Presentation of paper replications for arXiv Research codebases are notoriously difficult to set up We deploy auto-research agents to ingest popular arXiv repositories, resolve configuration issues and make the

→ View original post on X — @askalphaxiv

17 June 2026

New article by LeCun: visual learning via temporal differences

By

@askalphaxiv

–

17 June 2026 0h04

Another interesting article supervised by Yann LeCun! "You Don't Need Strong Assumptions: Visual Representation Learning via Temporal Differences" This article proposes Temporal Difference in Vision (TDV), which is a simple idea for learning vision from videos.

→ View original post on X — @askalphaxiv

17 June 2026

Introducing GLM-5.2 for Understanding Research Articles

By

@askalphaxiv

–

16 June 2026 23h30

Introducing GLM-5.2 for understanding research papers 🚀

Highlight any section of a paper to ask questions and “@” other papers for quick context, comparisons, and benchmark references pic.twitter.com/84LDzhLjhh
— alphaXiv (@askalphaxiv) 16 juin 2026

Highlight any section of an article to ask questions and '@' other articles for quick context, comparisons, and benchmark references.

→ View original post on X — @askalphaxiv

16 June 2026

Long video reasoning with DeepSeek sparse attention and multimodal MoE

By

@askalphaxiv

–

16 June 2026 19h57

"Kwai Keye-VL-2.0 Technical Report" This paper makes long-video reasoning much more feasible by adapting DeepSeek Sparse Attention to a GQA-based multimodal MoE, reaching 256K context with only 3B active parameters. As dense attention makes hour-level context way too expensive,

→ View original post on X — @askalphaxiv

16 June 2026

Accelerating RL Training via MTP and Rejection Sampling

By

@askalphaxiv

–

16 June 2026 6h12

Breaking entropy bounds: Accelerating RL training via MTP with rejection sampling. Given that RL training is limited by rollouts, and MTP is used to accelerate them, its acceptance rate would still collapse as the entropy of the

→ View original post on X — @askalphaxiv

16 June 2026

Dynamic Linear Attention for Adaptive Token Compression

By

@askalphaxiv

–

15 June 2026 23h45

"Dynamic Linear Attention" Most long-context linear models compress tokens using fixed blocks or logarithmic schedules, but long texts are not uniform. Stable segments can be summarized, while the

→ View original post on X — @askalphaxiv

15 June 2026

Mirage: Latent Spatial Memory for Video World Models

By

@askalphaxiv

–

15 June 2026 20h43

“Latent Spatial Memory for World Models” world models need a 3D memory to avoid drift, but RGB point cloud caches are slow because each step renders pixels and re-encodes them into latents via a VAE. The Mirage approach

→ View original post on X — @askalphaxiv

15 June 2026

Self-Evolving Multi-Agent Systems via Decentralized Memory

By

@askalphaxiv

–

15 June 2026 5h30

Multi-agent systems typically share a single memory pool, but this causes agents to converge towards the same behavior and leads to a loss of useful specialization. This article attributes to

→ View original post on X — @askalphaxiv

15 June 2026

Dense Supervision, Sparse Updates in On-Policy Distillation

By

@askalphaxiv

–

15 June 2026 5h30

"Dense Supervision, Sparse Updates: On the Sparsity and Geometry of On-Policy Distillation" OPD uses dense teacher feedback on student-generated rollouts, so we might expect dense parameter rewriting. However, this paper shows that is not the case. They found that OPD updates

→ View original post on X — @askalphaxiv

15 June 2026

From AGI to ASI: definition by Google DeepMind

By

@askalphaxiv

–

13 June 2026 18h43

This article by Google DeepMind defines how AGI is a unique human-level general system, and ASI is a system or collective that surpasses large expert human organizations in almost everything. They argue that the leap could

→ View original post on X — @askalphaxiv

13 June 2026