📄Papers

MV-Bench: Benchmarking MLLMs for Coordinated Multi-View Interface Construction

Zhao Y. et al.

2026-07

परिचय

जुलाई 2026 का benchmark, जो जाँचता है कि multimodal language models visual designs से executable multi-view visualization interfaces बना सकते हैं या नहीं। MV-Bench Tableau workbooks को structured ground truth की तरह इस्तेमाल करता है और visual fidelity, data bindings, coordinated interactions तथा code executability का मूल्यांकन करता है।

सारांश

MV-Bench क्या नया संकेत देता है:

1. Single-screen appearance से आगे: कई linked views वाले coordinated dashboards को test करता है
2. Executable benchmark: Code, rendered interfaces, datasets और interaction annotations के साथ 1,048 verified instances देता है
3. Structured ground truth: 92 Tableau workbooks से bindings और interactions निकालता है
4. Semantic evaluation: Visual fidelity के साथ data correctness और interaction completeness भी मापता है
5. महत्वपूर्ण capability gap: सबसे मजबूत tested model 75.45% layout accuracy तक पहुँचता है, लेकिन data-binding में केवल 21.71% और interaction accuracy में 11.68% हासिल करता है

मिलते-जुलते लेख

📄Papers

Toward Frontier-Quality Declarative UI Generation at Small-Model Cost

September 2026 Harness4GenUI paper studying whether small language models can generate production-oriented A2UI from approved component catalogs. Across task-management and cloud-console domains, the authors compare training-data strategies, model sizes, catalog sizes, cost, latency, binding correctness, and rendered quality.

a2uismall-language-modelssupervised-fine-tuning

लेखक Yang Y. et al. · 2026-09

EvoGenUI-Bench: Evaluating Multi-Turn Generative UI Assistants

August 2026 benchmark testing whether LLM assistants can maintain one executable web interface as user requirements evolve. EvoGenUI-Bench contains 150 five-turn tasks across information presentation, stateful interaction, and tool-grounded external state, with browser-based evaluation over screenshots, DOM and source evidence, actor traces, and runtime logs.

generative-uimulti-turnbenchmark

लेखक Peng Y. et al. · 2026-08

Maru: Information Architecture for Aligned, Persistent Generative UI

August 2026 HCI paper introducing information architecture as shared state for Generative UI. Maru captures how users partition, order, name, and prioritize information, turns those choices into persistent rules, and applies them across successive interface generations instead of rebuilding structure from each prompt.

generative-uiinformation-architecturepersonalization

लेखक Kim E. et al. · 2026-08

Awesome Generative UI