Papers by Mark Hasegawa-Johnson

10 papers
Finding Spoken Identifications: Using GPT-4 Annotation for an Efficient and Fast Dataset Creation Pipeline (2024.lrec-main)

Copied to clipboard

Challenge: Existing datasets that are limited to a few dialects, ethnicities, and age groups are not annotated considering these factors.
Approach: They propose a semi-automated dataset creation pipeline that leverages large language models to perform two complex annotation tasks using human annotations as ground truths.
Outcome: The proposed pipeline reduces time required for the filtering and tagging tasks while losing no important information.
Listen, Decipher and Sign: Toward Unsupervised Speech-to-Sign Language Recognition (2023.findings-acl)

Copied to clipboard

Challenge: Existing supervised sign language recognition systems rely on well-annotated data . instead, an unsupervised speech-to-sign language recognition system learns to translate between spoken and sign languages by observing only non-parallel speech and sign-language corpora.
Approach: They propose an unsupervised speech-to-sign language recognition system that can translate between spoken and sign languages by observing only non-parallel speech and sign-language corpora.
Outcome: The proposed approach outperforms baseline models on sign language corpora by 50% . the proposed approach is available at https://github.com/cactuswiththoughts/UnsupSpeech2Sign.git .
Self-supervised Semantic-driven Phoneme Discovery for Zero-resource Speech Recognition (2022.acl-long)

Copied to clipboard

Challenge: a new approach for self-supervised speech representation learning is proposed . a phoneme inventory learning model is based on a discrete representation of speech .
Approach: They propose a neural discrete representation learning model for self-supervised phoneme inventory learning with raw speech and word labels.
Outcome: The proposed model learns better phoneme-level representations and lowers error rates on TIMIT and Mboshi benchmarks than previous state-of-the-art models.
Worldly Wise (WoW) - Cross-Lingual Knowledge Fusion for Fact-based Visual Spoken-Question Answering (2021.naacl-main)

Copied to clipboard

Challenge: Question-Answering has long been of interest, but its accessibility to users through a speech interface and its support to multiple languages have not been addressed in prior studies.
Approach: They propose a task and a synthetically-generated dataset to do Fact-based Visual Spoken-Question Answering (FVSQA) the task requires a system to retrieve an entity from Knowledge Graphs (KGs) the question is spoken rather than typed.
Outcome: The proposed task performs at same levels of accuracy across 3 languages, including English, Hindi, and Turkish.
TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback (2024.findings-acl)

Copied to clipboard

Challenge: Existing approaches to provide token-level rewards fail to account for varying degrees of preference inherent to each token.
Approach: They propose a reward model that uses a discriminator to assign token-based continuous rewards to each token considering the context.
Outcome: Extensive experiments show that the proposed reward model improves on open-ended language generation benchmarks.
A Theory of Unsupervised Speech Recognition (2023.acl-long)

Copied to clipboard

Challenge: Unsupervised speech recognition (ASR) is the problem of learning automatic speech recognition systems from unpaired speech-only and text-only corpora.
Approach: They propose a general theoretical framework to study the properties of pasted macro ‘ASRU’/ systems based on random matrix theory and the theory of neural tangent kernels.
Outcome: The proposed framework proves various learnability conditions and sample complexity bounds on synthetic languages with three classes of transition graphs.
Syn2Vec: Synset Colexification Graphs for Lexical Semantic Similarity (2022.naacl-main)

Copied to clipboard

Challenge: In this paper we examine patterns of colexification as an aspect of lexical-semantic organization, and compare several approaches to build large scale graphs across 499 world languages.
Approach: They propose to use patterns of colexification as an aspect of lexical-semantic organization to build large scale synset graphs across a typologically diverse set of 499 world languages.
Outcome: The proposed models are evaluated against human judgments on a semantic similarity task for nine languages.
Query-based Cross-Modal Projector Bolstering Mamba Multimodal LLM (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing models that capture multiple modalities with a single input length are unable to handle this computational burden.
Approach: They propose a query-based cross-modal projector that compresses visual tokens based on input through the cross-attention mechanism.
Outcome: The proposed projector reduces the need for manually designing the 2D scan order of original image features when converting them into an input sequence for Mamba LLMs.
SMSMix: Sense-Maintained Sentence Mixup for Word Sense Disambiguation (2022.findings-emnlp)

Copied to clipboard

Challenge: Word Sense Disambiguation (WSD) is an NLP task aimed at determining the correct sense of a word in a sentence from discrete sense choices.
Approach: They propose a word-level mixup method that maintains the sense of a target word by increasing the frequency of least frequent senses during training to reduce the distributional bias of senses.
Outcome: The proposed method shows high accuracy on the most frequent sense (MFS) and low accuracy on least frequent senses (LFS).
INTapt: Information-Theoretic Adversarial Prompt Tuning for Enhanced Non-Native Speech Recognition (2023.findings-acl)

Copied to clipboard

Challenge: Existing approaches to improve ASR performance with pre-trained models require updating the pre-training model weights.
Approach: They propose a method that uses prompts concatenated to the original input to re-modulate attention of the pre-trained model.
Outcome: The proposed model improves the performance of L2 English and increases similarity between L2 and L1 accents.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations