In a paper to be presented at EMNLP 2024, we present a computational approach to efficiently model users’ latent susceptibility levels guided by the supervision of people's sharing behavior. The estimated susceptibility is significantly aligned with human judgments. This model enables large-scale susceptibility analysis for the first time.
Sep 19, 2024
Presenting three papers on bias, fairness and safety of Large Language Models at NAACL 2024 in Mexico City, including detecting and mitigating bias in QA models with ground-truth bias labels, fingerprinting LLMs, and a pilot study on injecting backdoors by instruction tuning data poisoning. Click for schedule and location details.
Jun 15, 2024
Presenting a demo and a poster at AAAI 2024 in Vancouver, including a demo on information diffusion via community-level information pathways and a poster on improving low-resource information extraction by structure-to-text data generation with Large Language Models.
Feb 22, 2024
Oral presentation of the conf paper: [Parameter-Efficient Low-Resource Dialogue State Tracking by Prompt Tuning](PromptDST). In the collaboration work with [Amazon Alexa AI](https://www.amazon.science/research-areas/conversational-ai-natural-language-processing), we introduce [a dialogue state tracking model](pubs/ma-etal-2023-parameter) tuning less than 1% of LM parameters and achieves better low-resource performance with prompt tuning techniques.
Aug 1, 2023
Presenting three papers at ACL 2023 in Toronto, including DICE: Data-Efficient Clinical Event Extraction with Generative Models, Multi-hop Evidence Retrieval for Cross-document Relation Extraction, and Can NLI Provide Proper Indirect Supervision for Low-resource Biomedical Relation Extraction? Click for schedule and location details.
Jul 1, 2023
New INTERSPEECH paper! In the collaboration work with [Amazon Alexa AI](https://www.amazon.science/research-areas/conversational-ai-natural-language-processing), we introduce [a dialogue state tracking model](../publication/ma-etal-2023-parameter) tuning less than 1% of LM parameters and achieves better low-resource performance with prompt tuning techniques.
May 1, 2023
I'll be at the [NeurIPS ENLSP](https://neurips2022-enlsp.github.io/) workshop.
Dec 1, 2022
In HyperVC, we use hyperbolic models with variable curvature to learn representations of temporal knowledge graphs for future fact prediction.
Sep 1, 2022
New EMNLP Findings paper. In HyperExpan, we use a more expressive hyperbolic space for taxonomy expansion.
Aug 1, 2021
We present our demo system EventPlus at NAACL 2021, the live system and code is released. [Check it out](https://kairos-event.isi.edu)!
Jun 1, 2021
The handbooks for ACL2020 that I helped assemble are ready and they're available in 24 timezones for the first time, [check them out](https://acl2020.org/participants/#conference-handbooks-for-24-timezones)!
Jul 1, 2020