A quick guide to Amazon’s 30+ papers on NAACL 2024

In recent years, the fields of natural language processing and computational linguistics, which were revolutionized a decade ago by deep learning, were again revolutionized by large language models (LLMs). It is not surprising that work involving LLMs, either as the subject of Infuny Therm, or as tools for other natural language processing applications, dominates at this year’s meeting in the North American chapter in Association for Computational Linguis (NAACL). This paper guide spells Amazon’s Naacl papers into those who are explicitly dealing with LLMs, and those, like you in many boxes, the ones you are present general techniques or data sets that can be used with either LLMs or more transitional models.

LLM-related work

Agents

Flap: Flow-sustained planning with limited decoding in llms
Shamik Roy, Sailik Sengupta, Daniele Bonadiman, Saab Mansour, Arshit Gupta

Autranty -Value Extraction

EIVEN: Effective implicitly allocated value extraction using multimodal llm
Henry Money Zou, Gavin Yu, Ziwei Fan, Dan Bu, Han Liu, Peng Dai, Dongmei Jia, Cornelia Caragea

Continuous learning

Q-Tuning: Tail-based fast tuning to lifelong get-shot language learning
Yanhui Guo, Shaoyuan Xu, Jinmiao Fu, Jia (Kevin) Liu, Chaosheng Dong, Bryan Wang

Dialogue

Utilization of LLMs for dialogue quality messages
Jinghan Jia, Abi Komom, Timothy Leffel, Xujun Money, Ajay Nagesh, Tamer Soliman, Aram Galstyan, Anoo Kumar

Mitigated hallucination

Less is more for improving automatic evaluation of ticketal consistency
Tong Wang, Ninad Kulkarni, Yanjun (Jane) Qi

Tofuval: Evaluation of Hallucinations of LLMs on Subject Focused Summary Dialogue
Liyan Tang, Igor Shalyminov, Amy Wong, Jon Burnsky, Jake Vincent, Yu’an Yang, Subi Singh, Song Feng, Hwanjun Song, Hang Su, Justin Sun, Yi Zhang, Saab Mansour, Kathleen McKkeown

Multi-Source enhanced Tignet Towaved to Long-Formed Response Generation
Nilay Patel, Shivasankar Subramanian, Siddhant Garg, Pratyay Banerjee, Amita Misra

Machine translation

A preference -driven paradigm for improved translation with large language models
Dawi Zhu, Sony Trenous, Xiaoyu Shen, Dietrich Klakow, Bill Byrne, Eva Hasler

Natural language treatment

Against informal language processing: Knowledge of slang in large language models
Zhewei Sun, Qian Hu, Rahul Gupta, Richard Zemel, Yang Xu

Answering questions

Bring your own kg: Self-monitored program synthesis to zero-shot kgqa
Dhruv Agarwal, Rajarshi (Raj) Das, Sopan Khosla, Rashmi Gangadharaiah

The universal question about answering questions that are ahead of “Bring Your Own KG: Self-Supervised Program Synthesis for Zero-Shot KGQA” uses a three-tight method to effectively adapt to new knowledge graph-without training data.

Reasoning

Comm
Pei Chen, Boran Han, Shuai Zhang

Recumend systems

Mind: Large language model -driven agent for recommendation
Yancheng Wang, Ziyan Jiang, Zheng Chen, Fan Yang, Yingxue Zhou, Eunah Cho, Xing Fan, Xiaojiang Huang, Yanbin Lu, Yingzhen Yang

Reinforcement learning from human feedback

RS-DPO: A sampling of hybrid rejection and direct preference optimization method for adjusting large language models
SAEED KHAKI, JINJIN LI, LAN MA, LIU YANG, PRATHAP RAMACHANDRA

AI Responsible

ITERALIGN: Constitutional iterative adaptation of large language models
Xusi chen, hongzhi wen, sreyashi nag, chen luo, qingyu yin, ruirui li, zheng li, wei wang

Mico: Preventive Detention of Large Language Models Through Inhibition Control
Roy Siegelmann, Ninareh Mehrabi, Palash Goyal, Prasoon Goyal, Lisa Bauer, Jwala Dhamala, Aram Galstyan, Rahul Gupta, Reza Ghanadan

The control of large language models against data -driven personas
Junyi Li, Charith Peris, Ninareh Mehrabi, Palash Goyal, Kai-Wei Chang, Aram Galstyan, Richard Zemel, Rahul Gupta

Performed Generated Generation

Improving contextual understanding in large language models through contrastive decoding
Zheng Zhao, Emilio Monti, Jens Lehmann, Haytham ASM

Text generation

Low cost generation and evaluation of dictionary examples phrases
Bill Cai, Clarence NG, Daniel Tan, Shelvia Hotama

Multi-Review Fusion-in-Context
Aviv Slobodkin, Ori Shapira, Ran Levy, Ido Dagan

Vision-language models

Magid: An automated pipeline for generating synthetic multimodal dataset
Hossein Aboutalebi, Justin Sun, Hwanjun Song, Yusheng Xie, Arshit Gupta, Hang SU, Igor Shalyminov, Nikolaos Pappas, Siffi Singh, Saab Mansour

Contributing to vision-language models for aspect-controlled generation of Réferring expressions
Danfeng Guo, Sanchit Agarwal, Arpit Gupta, Jiun-Yu Kao, Emre Barut, Tagyoung Chung, Jing Huang, Mohit Bansal

General and classic techniques

Conversation agents

Utilizing interesting facts to improve user engagement with conversation -boundary surfaces
Nikhita Vedula, Giuseppe Castellucci, Eugene Agichtein, Oleg Rokhlenko, Shervin Malmasi

Information extraction

Utilizing customer feedback for extraction of multimodal insight
Sandeep Sricharan Mukku, Abinesh Kanagarajan, PushPendu Ghosh, Chetan Aggarwal

Rexel: an end-to-end-model to extraction at document level
Nacime Bouziani, Shubhi Tyagi, Joseph Fisher, Jens Lehmann, Andrea Pierleoni

Machine learning

DRAW: Dynamic early exit on the decoder for accelerating coding decoder transformation models
Money Tang, Pengkai Zhu, Tian Li, Srikar Appalaraju, Vijay Mahadevan, R. Manmatha

Machine translation

Howl is bilingual lexicon induction?
Harsh Kohli, Helian Feng, Nicholas Dronen, Calvin McCarter, Sina Moeini, Ali Kebarighotbi

M3T: A new benchmark data set for multimodal document level machine translation
Benjamin HSU, Xiaoyu Liu, Huayang Li, Yoshinari Fujinuma, Maria Nedejde, Xing Niu, Yair Kittenplon, Ron Litman, Raghavendra Pappagari

AI Responsible

Limiting bias to questions about answering models by tracking bias influence
Mingyu Derek Ma, Jiun-Yu Kao, Arpit Gupta, Yu-Hsiang Lin, Wenbo Zhao, Tagyoung Chung, Wei Wang, Kai-Wei Chang, Nanyun Money

Semantic retrieval

Extremely effective online query that codes too tightly fetching
Nachshon Cohen, Yaron Fairstein, Guy Kushilevitz

Summary text

CCSUM: A large -scale dataset and high quality data set for abstract news summary
Xiang Jiang, Markus Dreyer

SEMI-mad abstract Dialogue Summary via high quality selection Pseudolabel
Jianfeng He, Hang SU, Jason Cai, Igor Shalyminov, Hwanjun Song, Saab Mansour

Visual question answer

Multiple-Questions Multiple-Svar Text-VQA
Peng Tang, Srikar Appalaraju, R. Manmatha, Yusheng Xie, Vijay Mahadevan

Leave a Comment