Challenge: Word-level alignment in speech-text pretraining models is limited by word-level annotated data . authors propose an iterative training method for USDP that reduces the dependency on scarce annotation resources.
Approach: They propose an Unsupervised Speech-text word-level alignment with Dynamic Programming (USDP) this method uses Dynamic programming principles to iteratively refine temporal alignment predictions .
Outcome: The proposed method significantly improves on speech-text pretraining tasks compared to existing methods.

Similar Papers

Speech-Text Pre-training for Spoken Dialog Understanding with Explicit Cross-Modal Alignment (2023.acl-long)

Copied to clipboard

Challenge: Existing speech-text pre-training methods are limited to one or two specific tasks, despite their success in speech-language processing tasks.
Approach: They propose a temporal position prediction task to capture the speech-text alignment . they use a textual dialog pre-training task to generalize a response selection task .
Outcome: The proposed model is superior in learning speech-text alignment and multi-turn dialog context.
CLASP: Cross-modal Alignment Using Pre-trained Unimodal Models (2024.findings-acl)

Copied to clipboard

Challenge: Recent advances in speech-text pretraining rely on parallel speech- text data . however, data accessibility is a challenge due to the limited data available.
Approach: They propose a framework for jointly performing speech and text processing without parallel corpora during pre-training but only downstream.
Outcome: The proposed framework extracts distinct representations for speech and text, aligning them effectively in a newly defined space using a multi-level contrastive learning mechanism.
WSPAlign: Word Alignment Pre-training via Large-Scale Weakly Supervised Span Prediction (2023.acl-long)

Copied to clipboard

Challenge: Existing word alignment methods rely on manual data and lack generalization ability.
Approach: They propose to use a weakly-supervised large-scale weakly supervised dataset for word alignment pre-training via span prediction to reduce the need for manual data.
Outcome: The proposed method improves upon the best supervised baseline by 3.3 6.1 points in F1 and 1.5 6.1 point in AER.
Multi-level Alignment Pretraining for Multi-lingual Semantic Parsing (2020.coling-main)

Copied to clipboard

Challenge: Existing methods for multilingual semantic parsing only handle monolingual parsers, while in real world applications such as Chatbot and search engine, we generally need to handle multi-lingual semanticparsing.
Approach: They propose a multi-level alignment pretraining method in a unified architecture for multi-lingual semantic parsing.
Outcome: The proposed method outperforms state-of-the-art methods on a publicly avail-able multi-lingual semantic parsing dataset and a newly constructed dataset.
SpanAlign: Sentence Alignment Method based on Cross-Language Span Prediction and ILP (2020.coling-main)

Copied to clipboard

Challenge: Existing methods for automatic sentence alignment assume monotonic alignments, but they can handle non-monotonic alignments.
Approach: They propose a method to automatically extract parallel sentences from noisy parallel documents by embeddings and encoding each source and target sentence.
Outcome: The proposed method improves translation accuracy by 4.1 BLEU scores on English-Japanese . it can predict spans in target document from sentences in source document .
Meeting the Needs of Low-Resource Languages: The Value of Automatic Alignments via Pretrained Models (2023.eacl-main)

Copied to clipboard

Challenge: Large multilingual models have inspired a new class of word alignment methods, which work well for pretraining languages.
Approach: They propose to use transformer-based word alignment methods to extract alignments from massive pretrained models.
Outcome: The proposed methods outperform traditional methods for languages unseen to pretraining models, and are competitive with each other.
Cross-Align: Modeling Deep Cross-lingual Interactions for Word Alignment (2022.emnlp-main)

Copied to clipboard

Challenge: Existing word alignment models capture few interactions between input sentence pairs, which severely degrades the word alignment quality.
Approach: They propose to model deep interactions between input and target sentences using a two-stage training framework to train the model.
Outcome: The proposed model achieves the state-of-the-art (SOTA) performance on four out of five language pairs.
Multi-Level Cross-Modal Alignment for Speech Relation Extraction (2024.emnlp-main)

Copied to clipboard

Challenge: Existing studies use synthetic speech to train and evaluate SpeechRE models, hindering their development . modality gap issue limits performance of existing models, limiting future researches .
Approach: They propose to use speech data to train and evaluate SpeechRE models by using real speech . they propose to train a cross-modal alignment model to bridge the modality gap .
Outcome: The proposed model can train to bridge the modality gap between speech encoder and text decoder . the proposed model is based on two real SpeechRE datasets .
Improving Neural Language Models by Segmenting, Attending, and Predicting the Future (P19-1)

Copied to clipboard

Challenge: Common language models typically predict the next word given a past context.
Approach: They propose a method that aligns the given context and the following phrase . they define syntactic heights and phrase segmentation rules to enable it to learn .
Outcome: The proposed model outperforms strong baseline models on Wikitext-103 dataset.
Cross-lingual Spoken Language Understanding with Regularized Representation Alignment (2020.emnlp-main)

Copied to clipboard

Challenge: despite promising results, current cross-lingual models suffer from imperfect cross-linguistic representation alignments between the source and target languages, which makes the performance sub-optimal.
Approach: They propose a regularization approach to align word-level and sentence-level representations across languages without external resources.
Outcome: The proposed model outperforms state-of-the-art models in few-shot and zero-shot scenarios and achieves comparable performance to supervised training with all training data.

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