Akari Asai

@[email protected]
171 Followers
85 Following
30 Posts

NLP Ph.D. student at University of Washignton Computer Science and Engineering.

IBM PhD fellow (2022-).

Currently:
UW NLP & Part-time at Meta AI.

Formely:
EECS undergrad at UTokyo
2x Engineering internships at Google
Research internships at Microsoft Research Asia, Megagon lab and salesforce research

In my free time, I πŸƒπŸ»β€β™€οΈπŸ‘©πŸ»β€πŸ³πŸ₯§πŸ§—πŸ»β€β™€οΈπŸ“–πŸ‘©πŸ»β€πŸ’»

Websitehttps://akariasai.github.io/
Twitter@AkariAsai
LocationSeattle, WA
Pronounsshe/her
Akari AsaiDec 20, 2022

New paper 🚨

Can we solely rely on LLMs’ memories (eg replace search w ChatGPT)? Probably not.
Is retrieval a silver bullet? Probably not either.

Our analysis reveals that LLMs' memorizations are still limited and scaling won't help much in long-tail distributions.
We show that adaptively incorporating non-parametric memories (eg retrieved chunks) can improve performance as well as efficiency.

πŸ“œ http://tinyurl.com/2sdeuupn πŸ’» http://github.com/AlexTMallen/adaptive-retrieval

#PaperThread #newpaper
[1/N]

Show threadAkari AsaiDec 20, 2022
LLMs store a lot of factual knowledge in their parameters (parametric factual knowledge), but recent work shows that they struggle to learn less frequent facts and can often hallucinate when they don't know. How much do they memorize and what affects their memorization? [2/N]
Akari AsaiDec 19, 2022
A deadline life pro tip: now you can use Grammarly with Overleaf ✍️
https://www.overleaf.com/learn/how-to/Use_Grammarly_with_Overleaf
Use Grammarly with Overleaf

An online LaTeX editor that’s easy to use. No installation, real-time collaboration, version control, hundreds of LaTeX templates, and more.

Akari AsaiDec 19, 2022
Akari AsaiNov 6, 2022

#Introduction I am a 4th-year #NLProc Ph.D. student at University of Washington.
I'm broadly interested in how to build NLP models that work well on many tasks (e.g., a wide range of tasks/tasks requiring complex reasoning) and languages, with a focus on knowledge-intensive NLP tasks (e.g., Question Answering).
Before Ph.D., I completed my EECS Bachelor at the University of Tokyo πŸ—Ό.
I'm currently doing my part-time at Meta AI Seattle.

https://akariasai.github.io/

Akari Asai

Akari Asai
Akari AsaiNov 29, 2022

As a life-long Harry Potter fan, I'm glad to see more and more papers get their title inspirations from "Fantastic Beasts and Where to Find Them", although I I have very mixed feelings about the most recent film 🌢️ ...

https://aclanthology.org/2022.acl-long.34/
https://aclanthology.org/2022.acl-long.556/

Fantastic Questions and Where to Find Them: FairytaleQA – An Authentic Dataset for Narrative Comprehension

Ying Xu, Dakuo Wang, Mo Yu, Daniel Ritchie, Bingsheng Yao, Tongshuang Wu, Zheng Zhang, Toby Li, Nora Bradford, Branda Sun, Tran Hoang, Yisi Sang, Yufang Hou, Xiaojuan Ma, Diyi Yang, Nanyun Peng, Zhou Yu, Mark Warschauer. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

ACL Anthology
Akari AsaiNov 18, 2022
Anjalie FieldNov 16, 2022
I will be looking for 1-2 PhD students this cycle in social/ethical NLP, more information here: https://anjalief.github.io/research_areas.html
Anjalie Field, CMU

Akari AsaiNov 18, 2022
We further introduce a new and realistic setup, cross-task cross-domain evaluation, where queries with diverse intents and documents are all pooled and queries only don't fully capture intents. TART largely outperforms competitive models on this setup as well.
7/N
Show threadAkari AsaiNov 18, 2022
A user query can have diverse intents (e.g., retrieve relevant documents, find code implementation or similar questions asked in the forum previously).
We often build separate retrieval systems for different intents by training a retriever to model those implicit intents. 2/N
Show threadAkari AsaiNov 18, 2022

We advocate for a new task formulation, retrieval with Instructions, where a retriever takes a query AND an instruction that EXPLICITLY describes the information need.

The goal here is to build a single retriever that can find relevant documents satisfying the instruction. 3/N

Akari AsaiNov 18, 2022

New paper 🚨 https://arxiv.org/abs/2211.09260

Can we train a single search system that satisfies our diverse information needs?

We present 𝕋𝔸ℝ𝕋 πŸ₯§ the first multi-task instruction-following retriever trained on 𝔹𝔼ℝℝ𝕀 🫐, a collections of 40 retrieval tasks with instructions! 1/N

#PaperThread #newpaper

Task-aware Retrieval with Instructions

We study the problem of retrieval with instructions, where users of a retrieval system explicitly describe their intent along with their queries, making the system task-aware. We aim to develop a general-purpose task-aware retrieval systems using multi-task instruction tuning that can follow human-written instructions to find the best documents for a given query. To this end, we introduce the first large-scale collection of approximately 40 retrieval datasets with instructions, and present TART, a multi-task retrieval system trained on the diverse retrieval tasks with instructions. TART shows strong capabilities to adapt to a new task via instructions and advances the state of the art on two zero-shot retrieval benchmarks, BEIR and LOTTE, outperforming models up to three times larger. We further introduce a new evaluation setup to better reflect real-world scenarios, pooling diverse documents and tasks. In this setup, TART significantly outperforms competitive baselines, further demonstrating the effectiveness of guiding retrieval with instructions.

arXiv.org