So you want to ship an AI product? New blog post. austinhenley.com/blog/copilotpa…

We've made a dataset of ~7k student queries from CodeAid publicly available: huggingface.co/datasets/majee… 📊 Example use cases: - Automatically categorize student queries - Benchmark educational AI tools against real-world queries - Study the role and effectiveness of AI in CSEd

How do AI code generators effect students? New blog post with Majeed Kazemi summarizing our recent research. austinhenley.com/blog/learningw…

We introduce Sketch-Then-Generate — a new way to code with #LLM. It offers instant, *incremental* feedback via code sketches as you type prompts, and uses these sketches as outlines to guide LLMs in code generation arxiv.org/abs/2405.03998 w/ Zeyu, Xiaoshuo, and Elena Glassman

Human-Computer Interaction research from U of T Department of Computer Science is well represented at #CHI2024 with over 30 accepted papers, including a Best Paper Award and a few honourable mentions! uoft.me/au0

Need something to read this weekend? Take a look at our UIST paper!

HCI systems papers need to tell visual stories. As we near the CHI deadline, here's a set of exemplary figures curated by my lab. Hope it helps inspire! Feel free to contribute some examples. docs.google.com/presentation/d…

So cool to see OpenAI integrating Chain-of-Thought into its o1 models. Our recent #UIST2024 paper is all about making the AI's chain-of-thought interactive and editable! 🤖 > 🧠 > 📝 > 🏗️ 📄arxiv.org/abs/2407.02651

Done with writing for #CHI2025 - how about some reading for #UIST2024? 😄 Here's a collection of preprints. Suggestions welcome! medium.com/@dbuschek/uist… Justin Weisz Werner Geyer Elizabeth Anne Watkins, PhD Qiaosi (Chelsea) Wang Alan Said

As we work on those #CHI2025 videos 🎥 (deadline: Sept 19), here is a set of inspirational videos curated by my lab. Would love more examples, especially for other HCI research contribution types. Send them my way! docs.google.com/presentation/d…

Excited to present our paper at #UIST2024 tomorrow in the "AI as Copilot" session (4:10 PM) I'll be discussing how breaking down data analysis tasks into *editable*, *progressively revealed* steps can enhance control over AI and improve verification. 📜 dl.acm.org/doi/10.1145/36…