Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userSequence Model with Self-Adaptive Sliding Window for Efficient Spoken Document Segmentation
71
0
0.0
(
0
)
Added by
Qinglin Zhang
Publication date
2021
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
Transcripts generated by automatic speech recognition (ASR) systems for spoken documents lack structural annotations such as paragraphs, significantly reducing their readability. Automatically predicting paragraph segmentation for spoken documents may both improve readability and downstream NLP performance such as summarization and machine reading comprehension. We propose a sequence model with self-adaptive sliding window for accurate and efficient paragraph segmentation. We also propose an approach to exploit phonetic information, which significantly improves robustness of spoken document segmentation to ASR errors. Evaluations are conducted on the English Wiki-727K document segmentation benchmark, a Chinese Wikipedia-based document segmentation dataset we created, and an in-house Chinese spoken document dataset. Our proposed model outperforms the state-of-the-art (SOTA) model based on the same BERT-Base, increasing segmentation F1 on the English benchmark by 4.2 points and on Chinese datasets by 4.3-10.1 points, while reducing inference time to less than 1/6 of inference time of the current SOTA.
rate research
Read More
Recently abstractive spoken language summarization raises emerging research interest, and neural sequence-to-sequence approaches have brought significant performance improvement. However, summarizing long meeting transcripts remains challenging. Due to the large length of source contents and targeted summaries, neural models are prone to be distracted on the context, and produce summaries with degraded quality. Moreover, pre-trained language models with input length limitations cannot be readily applied to long sequences. In this work, we first analyze the linguistic characteristics of meeting transcripts on a representative corpus, and find that the sentences comprising the summary correlate with the meeting agenda. Based on this observation, we propose a dynamic sliding window strategy for meeting summarization. Experimental results show that performance benefit from the proposed method, and outputs obtain higher factual consistency than the base model.
Sequence labeling is an important technique employed for many Natural Language Processing (NLP) tasks, such as Named Entity Recognition (NER), slot tagging for dialog systems and semantic parsing. Large-scale pre-trained language models obtain very good performance on these tasks when fine-tuned on large amounts of task-specific labeled data. However, such large-scale labeled datasets are difficult to obtain for several tasks and domains due to the high cost of human annotation as well as privacy and data access constraints for sensitive user applications. This is exacerbated for sequence labeling tasks requiring such annotations at token-level. In this work, we develop techniques to address the label scarcity challenge for neural sequence labeling models. Specifically, we develop self-training and meta-learning techniques for training neural sequence taggers with few labels. While self-training serves as an effective mechanism to learn from large amounts of unlabeled data -- meta-learning helps in adaptive sample re-weighting to mitigate error propagation from noisy pseudo-labels. Extensive experiments on six benchmark datasets including two for massive multilingual NER and four slot tagging datasets for task-oriented dialog systems demonstrate the effectiveness of our method. With only 10 labeled examples for each class for each task, our method obtains 10% improvement over state-of-the-art systems demonstrating its effectiveness for the low-resource setting.
This work proposes a novel adaptation of a pretrained sequence-to-sequence model to the task of document ranking. Our approach is fundamentally different from a commonly-adopted classification-based formulation of ranking, based on encoder-only pretrained transformer architectures such as BERT. We show how a sequence-to-sequence model can be trained to generate relevance labels as target words, and how the underlying logits of these target words can be interpreted as relevance probabilities for ranking. On the popular MS MARCO passage ranking task, experimental results show that our approach is at least on par with previous classification-based models and can surpass them with larger, more-recent models. On the test collection from the TREC 2004 Robust Track, we demonstrate a zero-shot transfer-based approach that outperforms previous state-of-the-art models requiring in-dataset cross-validation. Furthermore, we find that our approach significantly outperforms an encoder-only model in a data-poor regime (i.e., with few training examples). We investigate this observation further by varying target words to probe the models use of latent knowledge.
The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with Hepos, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GovReport, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.
The scheme of the sliding window is known in Information Theory, Computer Science, the problem of predicting and in stastistics. Let a source with unknown statistics generate some word $... x_{-1}x_{0}x_{1}x_{2}...$ in some alphabet $A$. For every moment $t, t=... $ $-1, 0, 1, ...$, one stores the word (window) $ x_{t-w} x_{t-w+1}... x_{t-1}$ where $w$,$w geq 1$, is called window length. In the theory of universal coding, the code of the $x_{t}$ depends on source ststistics estimated by the window, in the problem of predicting, each letter $x_{t}$ is predicted using information of the window, etc. After that the letter $x_{t}$ is included in the window on the right, while $x_{t-w}$ is removed from the window. It is the sliding window scheme. This scheme has two merits: it allows one i) to estimate the source statistics quite precisely and ii) to adapt the code in case of a change in the source statistics. However this scheme has a defect, namely, the necessity to store the window (i.e. the word $x_{t-w}... x_{t-1})$ which needs a large memory size for large $w$. A new scheme named the Imaginary Sliding Window (ISW) is constructed. The gist of this scheme is that not the last element $x_{t-w}$ but rather a random one is removed from the window. This allows one to retain both merits of the sliding window as well as the possibility of not storing the window and thus significantly decreasing the memory size.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
