ricardo

MPhil student @HKUST(GZ)

[Han Peng @hahapaper]

Our work is featured in Hugging Face Daily Papers!

RULER: Instance-aware Rubric Rewards for SVG Generation. https://t.co/v39zdvwxZF

Huge credit to our student authors Hangyu Ran @zzranhangyu, Yuhao Zheng @xinlinzyh, and our collaborator Kevin Qinghong Lin @KevinQHLin, for their excellent work!

Building on our experience with large-scale post-training, I’ve been thinking about three practical challenges in pushing model capabilities further:

• How do we make open-ended tasks meaningfully verifiable?
• How do we align rewards with human taste and aesthetic judgment?
• How do we make feedback detailed and task-specific, yet scalable?

Our work takes a concrete step in this direction: RULER derives evaluation criteria from each user prompt, assesses the rendered SVG item by item, and turns those judgments into RL rewards.

[DailyPapers @HuggingPapers]

RULER: Rubric Rewards for SVG Generation

Ant Group researchers introduce a rubric-based RL reward that turns each prompt into a six-item, instance-aware rubric, scores rendered SVGs with a judge VLM, and needs no SVG ground truth or human preference labels.

[David K 🎹 @DavidKPiano]

I'd be really surprised if AI ever "one-shots" something like this using just HTML + CSS

[Kevin Lin @KevinQHLin]

We added Gemini3-Pro #Gemini results on VCode!
Check out at https://t.co/YPIvEmySnB

🏅Gemini3-Pro performs quite well at image2svg.
It achieved 52.1 in VCode, surpassing GPT-5's 46.8 (+5.3).

Open-sourced at https://t.co/2R7F42rOp6

[Simon Willison @simonw]

Playing with a new LLM benchmark: how good are they at drawing an SVG of a pelican on a bicycle?

Here's the difference between Claude 3.5 Sonnet new and Claude 3.5 Sonnet previous

[Jinpeng Wang @awinyimgprocess]

Unique avatar 🤣🤣🤣😂😂😂
Try demo at: https://t.co/3ZFWuuZP34

[AK @_akhaliq]

app: https://t.co/ei0thfiS6h

[AK @_akhaliq]

VCode app is out

SVG as Symbolic Visual Representation

Turn any image into symbolic SVG

[Jinpeng Wang @awinyimgprocess]

Turn any interesting imgs into svg🤣🤣🤣🤣🤣

[Kevin Lin @KevinQHLin]

😲Turn any image into symbolic SVG at
https://t.co/xFyd3qgJaT
💻Github: https://t.co/yDscg1r3jj

Built with @xinlinzyh and @zzranhangyu
Thanks @_akhaliq for sharing and Gradio demo team
@huggingface @Gradio!

[AK @_akhaliq]

VCode

a Multimodal Coding Benchmark with SVG as Symbolic Visual Representation

[DailyPapers @HuggingPapers]

Microsoft Research unveils VCode

A new multimodal coding benchmark that turns images into expressive SVG code for advanced visual reasoning. It's time to bridge the gap between pixel-based vision and symbolic understanding!

[DailyPapers @HuggingPapers]

Check out VCode, a benchmark redefining visual understanding as code generation, and VCoder, an agentic framework achieving +12.3 pts over SOTA!

Paper: https://t.co/r6D4KA5SAZ