encounter1997

A phd student majoring in Computer Vision.

[encounter1997 @encounter19972]

Full details and theoretical proofs are in the paper. Happy to discuss and answer questions.

[encounter1997 @encounter19972]

Sharing our latest work (StreamMA): exploring how to make multi-agent reasoning faster, more accurate, and cheaper. Hop on HuggingFace for an upvote, GitHub for a star 🤗:

HuggingFace: https://t.co/BgajgGFziE

Project: https://t.co/5qWvp3o1Ek

GitHub: https://t.co/sZrXjL97CT

[encounter1997 @encounter19972]

Hard numbers: ① +7.3pp avg over 8 benchmarks across math / science / code (Claude Opus 4.6-high); ② 26.9× wall-clock speedup at A=64, S=64 (83% of the theoretical bound); ③ Stream×4 at half the price ($2.75 vs $5.46) beats Serial×16 (90.9% vs 89.4%).

[AK @_akhaliq]

The World is Your Canvas

Painting Promptable Events with Reference Images, Trajectories, and Text

[AK @_akhaliq]

MagicQuillV2

Precise and Interactive Image Editing with Layered Visual Cues

[AK @_akhaliq]

Meta presents TUNA

Taming Unified Visual Representations for Native Unified Multimodal Models

[weijia wu @weijiawu7]

🔥 New paper out: WEAVE — a 100K-sample interleaved multimodal dataset + WEAVEBench, a human-annotated benchmark for visual memory, multi-turn editing.
📄 arXiv: https://t.co/wAE9Wvy7xy
🐙 GitHub: https://t.co/6roLGE4CmZ
🤗 HF Dataset: https://t.co/gTrygOAyne

[AK @_akhaliq]

HoloCine

Holistic Generation of Cinematic Multi-Shot Long Video Narratives

[Bruce Yue Yu @Bruce_YuYue]

🤩🎬 HoloCine is here!

The first open-source multi-shot long video model, generating minute-long cinematic narratives as stunning as Sora 2.

Watch the demo ↓

[DogeDesigner @cb_doge]

Elon Musk v/s Sam Altman

The A.I. simulation got way too real.

[Yongliang Shen @itricktreat]

Introducing GSM8K-V: Can vision-language models solve grade school math when problems are shown visually instead of text? 🧮👁️
We converted all 1,319 GSM8K problems into comic-style multi-image sequences (5,343 images total).
The results are surprising! 🧵

[Yongliang Shen @itricktreat]

Introducing EasySteer: High-performance LLM steering framework built on vLLM. Achieves 5.5-11.4× speedup over existing tools while maintaining 71-84% throughput.
Paper: https://t.co/KtQFs54FuH
Code: https://t.co/Oyrk8QG1r8
HF Paper: https://t.co/aS4BwXY8Xp

[机器之心 JIQIZHIXIN @jiqizhixin]

What if LLMs already had the right answer—but erased it before finishing? 🤯

New work on diffusion LLMs (dLLMs) uncovers temporal oscillation: correct answers often appear mid-denoising, only to vanish in later steps.

Two fixes that harness temporal consistency:

- Temporal Self-Consistency Voting → training-free decoding that aggregates stable predictions across steps
- Temporal Consistency Reinforcement → post-training with Temporal Semantic Entropy (TSE) as a reward for semantic stability

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models

"Our work here reveals a critical phenomenon, temporal oscillation, where correct answers often emerge in the middle process, but are overwritten in later denoising steps. To address this issue, we introduce two complementary methods that exploit temporal consistency: 1) Temporal Self-Consistency Voting, a training-free, test-time decoding strategy that aggregates predictions across denoising steps to select the most consistent output; and 2) a post-training method termed Temporal Consistency Reinforcement, which uses Temporal Semantic Entropy (TSE), a measure of semantic stability across intermediate predictions, as a reward signal to encourage stable generations."