Life Update: I’m excited to announce I’ve joined @SFResearch! With that, I’ll be sharing some of my final papers with Adobe Research at #UIST2025: 🌅 Thematic Planes: Bridging Tacit User Intent and Latent Spaces for Image Generation w/ @nikhilsksharma, @donghoon_io, @daeun_choi_, Harsh Sharma, @MasterJeongK, @hengjinlp Link: arxiv.org/abs/2508.06065 🧑🏻‍💻 PosterMate: Audience-driven Collaborative Persona Agents for Poster Design w/ @donghoon_io, @garyhsieh, @GromitC Link: arxiv.org/abs/2507.18572 📝 Semantic Commit: Helping Users Update Intent Specifications for AI Memory at Scale w/ @priyanvaithi, @Kim_Munyeong, Frida-Cecilia Acosta-Parenteau, Amine Mhedhbi, @roboticwrestler, @IanArawjo Link: arxiv.org/abs/2504.09283 *Special thanks to Ian and Priyan for having me :) I had an amazing time with Adobe Research, where I met truly kind and brilliant mentors/collaborators. Come chat with some of them at these posters!

📣 @Adobe is at #CVPR2025! Come chat with me on our latest projects! 👋🏻 You can catch me at the booth, or our demo on Saturday from 5 PM - 7 PM at Exhibit Hall D on factual multimodal generations!

📣 Attending @RealAAAI? Come check out our paper on understanding parametric knowledge in text-to-video models, and how external knowledge can improve factuality! 📄 We focus on knowledge-intensive prompts containing entities, and implement a lightweight query distiller knowledge retriever for an initial investigation. Watch out for an expanded version at a future conference! Location: Pennsylvania Convention Center, Exhibit Hall E, Level 2 @ Feb 27 - 12:30 PM ⭐️ Had a great time collaborating with @yunyao_li and Arjun Chandra from @Adobe, Yang Zhao from @AdobeResearch and @ConiaSimone from @SapienzaRoma. 🔗 Link to the paper is coming soon. Check out our demo video (bit.ly/factv-demo) and GitHub repo (bit.ly/factv-dataset) in the meantime!

☕️ A bittersweet paper announcement at @coling2025! I’m excited to announce a couple of more papers from @Apple and @Adobe. 🌎 KG-TRICK: Unifying Textual and Relational Information Completion of Knowledge for Multilingual Knowledge Graphs @ Main Track w/ Zelin Zhou, @ConiaSimone, Min Li, Shenglei Huang, Umar Farooq Minhas, @saloni_p, Henry Xiao, @yunyao_li 🌄 MuRAR: A Simple and Effective Multimodal Retrieval and Answer Refinement Framework for Multimodal Question Answering @ Demo Track w/ Zhengyuan Zhu, Hong Zhang, Sai Harsha, Loic Feujio, Akash Maharaj, @yunyao_li 🌟 Thank you to all the authors, but a special thank you to Simone! KG-TRICK is the last of a series of 4 papers with Simone on multilingual and cross-lingual semantics, where he (and @gangi_official @rpradeep42) at Apple introduced and guided me on how to conduct NLP research! Excited to share more work in the coming months, where they’ve continued to inspire me.

🎉 Our paper won an Outstanding Paper Award at #EMNLP2024! 🌟 “Towards Cross-Cultural Machine Translation with Retrieval-Augmented Generation from Multilingual Knowledge Graphs” was recognized for addressing “the challenge of translating culturally-relevant named entities.” 📣 A huge thank you to the brilliant people behind this project, including @ConiaSimone, Min Li, Umar Farooq Minhas, @saloni_p and @yunyao_li. 📄: aclanthology.org/2024.emnlp-… 💻: github.com/apple/ml-kg-mt 👥 As a junior researcher, I’m extremely grateful and fortunate to have worked with and been mentored by incredible researchers like: @ConiaSimone (@SapienzaNLP), @gangi_official (@IllinoisCDS), @rpradeep42 (@UWaterloo), @saloni_p (@Apple), @yunyao_li (@Adobe) and more. Keep an eye out for more work from us in the coming months!

🏖️I’ll be in Miami till the 16th for #EMNLP2024 presenting ConvKGYarn 🧶 in the industry track session E on the 13th, 4-5:30 (Riverfront) with @daniel_js_lee! Happy to chat about eval in @TREC_RAG (& convince you to join next year!), rerankers or whatever else has you curious.

📣 Excited to announce 3 papers were accepted at #EMNLP2024 from my time at @Apple and @Adobe! 📄 Towards Cross-Cultural Machine Translation with Retrieval-Augmented Generation from Multilingual Knowledge Graphs @ Main Track w/ @ConiaSimone, Min, @umarfm13, @saloni_p, @yunyao_li 📄 ConvKGYarn: Spinning Configurable and Scalable Conversational Knowledge Graph QA Datasets with Large Language Models @ Industry Track w/ @rpradeep42, Ali, Jeff, @YisiSang, @lintool, @ihabilyas, @saloni_p, Mostafa, @yunyao_li 📄 RETAIN: Interactive Tool for Regression Testing Guided LLM Migration @ Demo Track w/ @tanayd07, Sally, Sai, Anirudh, Akash, @yunyao_li 🌟A huge thank you to all the authors, especially Akash, Min, Mostafa, @saloni_p and @yunyao_li on advising on these projects! 🌎 We will see you in Miami in November! (Or at Disney World in Orlando…) Read more about the papers below ⬇️

Come check out our recent paper from @Apple! @rpradeep42 is presenting ConvKGYarn at #ACL2024 today at @kallm_ACL2024 (4:45 PM) and tomorrow at @lm_knowledge (4:00 PM) in Convention Center A1!