Current AI Research Scientist at Intel Lab, PhD from ASU, former intern @Google @Meta @Salesforce, working on natural language processing and vision+language.
Joined April 2020
- Tweets48
- Following270
- Followers131
- Likes138
Thrilled to share our ICML oral! 🎉 We release a large-scale, diverse synthetic multimodal dataset for better knowledge base VQA (RAG)—balanced diversity & quality. Come to talk with Phillip and Kris🌟 #ICML2025 #SyntheticData #Multimodal
2
190
The talk given by @polynoamial at the MATH workshop #NeurIPS2024 is very inspiring: the inference scaling seems to offer a trade-off compared to training scaling laws. Additionally, ‘LLM Verifier’ is a way to scale up inference time. What is your favorite verifier technique?
2
2
268
Multimodal4healthcare workshop @ICHI2024 will start at 8:30AM (EDT) on June 3rd at Room 334 at UF Research and Academic Center at Lake Lona. Welcome to join and interact with amazing speakers @james_y_zou @pengyifan @aaronylee and five selected paper presenters. Looking forward!
Our deadline is extended to March 8th! Looking forward to your submission !!!
We're thrilled to share that our Multimodal4Health workshop will take place during the ICHI 2024 conference. Our "Call for Papers" is now open: luomancs.github.io/Multimoda… ! We are looking forward to seeing your innovative and thought-provoking work!
3
440
If you are working on research related to multimodal for health, we warmly invite you to submit your work to our workshop at the upcoming ICHI conference! It’s a good opportunity to meet and chat with some of the related researcher in the field, get feedback!
We're thrilled to share that our Multimodal4Health workshop will take place during the ICHI 2024 conference. Our "Call for Papers" is now open: luomancs.github.io/Multimoda… ! We are looking forward to seeing your innovative and thought-provoking work!
1
243
We're thrilled to share that our Multimodal4Health workshop will take place during the ICHI 2024 conference. Our "Call for Papers" is now open: luomancs.github.io/Multimoda… ! We are looking forward to seeing your innovative and thought-provoking work!
The event will feature three presentations by prominent researchers in the field, along with four papers chosen for oral presentations. Please help us by spreading the word to those who might be interested! Organizers: @ImonBanerjee6 @AishaUroojKhan @bhavik_md @jasonafries
(1/2)Retrieval In-context learning(Ret-ICL) stands out by using a low-cost retriever to select highly relevant demonstrations for each query, offering a more efficient alternative to full model or efficient-parameter fine-tuning (EPFT). Want to know more? arxiv.org/pdf/2401.11624.pdf
I was unable to attend the ACL Conference in person due to visa issue. Hope all those present have an enjoyable and inspiring experience! We have two retrieval (multimodal and text domain) papers accepted to #ACL2023NLP main conference, Check them out :)
1
5
409
1. arxiv.org/abs/2306.00424. We introduce a new multimodal-query retrieval benchmark with an end-to-end multimodal retriever, ReMuQ dataset is available online: github.com/luomancs/ReMuQ. It is a collaboration with @Jacob292020 @trgokhale @Yezhou_Yang @chittabaral
2. arxiv.org/pdf/2210.01371.pdf studies two important aspects of text retriever: indexing efficiency and generalization. One important finding is that there is a large room for improving the robustness of retriever. Thanks the great feedback from my Meta internship team!
133
@geoffreyhinton
just gave a keynote talk at #ACL2023NLP, It was a great talk and here are few aspects that inspires me: hallucination, mortal computing, knowledge distillation, knowledge sharing, subjective experience. "I like the problem that everybody knows that you are wrong."
1
1
190
Knowledge Sharing: my summary from the talk is that this can be achieved via two ways, one is between multiple agents, the other one is super intelligent model that learn from internet through video and audio.
1
97