Papers by David Vazquez

4 papers
StarFlow: Generating Structured Workflow Outputs From Sketch Images (2026.eacl-long)

Copied to clipboard

Challenge: Despite being widely used, building workflows can be complex, often requiring manual configuration through low-code platforms or visual programming tools.
Approach: They propose a framework for generating structured workflow outputs from sketches using vision-language models to automate the process.
Outcome: The proposed framework outperforms large vision-language models in the task of generating structured workflow outputs from sketches and diagrams.
XC-Cache: Cross-Attending to Cached Context for Efficient LLM Inference (2024.findings-emnlp)

Copied to clipboard

Challenge: XC-Llama uses pre-trained decoder-only models to condition generation on reference text without the prompt.
Approach: They propose a model that uses cross-attention to condition generation on reference text without the prompt.
Outcome: The proposed models outperform prompt-based inference methods and reduce space footprint relative to standard KV caching by two orders of magnitude.
TK-KNN: A Balanced Distance-Based Pseudo Labeling Approach for Semi-Supervised Intent Classification (2023.findings-emnlp)

Copied to clipboard

Challenge: Semi-supervised methods for detecting intent generate a large amount of unlabeled data . labeling data requires substantial human effort, and picking an imbalanced set of examples could lead to poor labels.
Approach: They propose a balanced distance-based pseudo-labeling approach for semisupervised intent classification . they use a ranking-based approach to select samples with a model prediction confidence .
Outcome: The proposed method outperforms existing models on popular datasets.
WebMMU: A Benchmark for Multimodal Multilingual Website Understanding and Code Generation (2025.emnlp-main)

Copied to clipboard

Challenge: Existing benchmarks focus on specific aspects of web tasks but lack comprehensive coverage.
Approach: They propose a multilingual benchmark that evaluates three core web tasks: (1) website visual question answering, (2) code editing involving HTML/CSS/JavaScript, and (3) mockup-to-code generation.
Outcome: The proposed model performs well on basic information extraction, but struggles with reasoning and grounding, editing code to preserve functionality, and generating design-to-code that maintains hierarchy and supports multilingual content.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations