Unsupervised Learning of Discourse Structures using a Tree Autoencoder
نویسندگان
چکیده
Discourse information, as postulated by popular discourse theories, such RST and PDTB, has been shown to improve an increasing number of downstream NLP tasks, showing positive effects synergies with important real-world applications. While methods for incorporating become more sophisticated, the growing need robust general structures not sufficiently met current parsers, usually trained on small scale datasets in a strictly limited domains. This makes prediction arbitrary tasks noisy unreliable. The overall resulting lack high-quality, high-quantity trees poses severe limitation further progress. In order alleviate this shortcoming, we propose new strategy generate tree task-agnostic, unsupervised fashion extending latent induction framework auto-encoding objective. proposed approach can be applied any tree-structured objective, syntactic parsing, parsing others. However, due especially difficult annotation process trees, initially develop method larger diverse treebanks. paper are inferring natural text multiple domains, promising results set tasks.
منابع مشابه
Unsupervised Learning of Task-Specific Tree Structures with Tree-LSTMs
For years, recursive neural networks (RvNNs) have been shown to be suitable for representing text into fixed-length vectors and achieved good performance on several natural language processing tasks. However, the main drawback of RvNNs is that they require structured input, which makes data preparation and model implementation hard. In this paper, we propose Gumbel Tree-LSTM, a novel treestruct...
متن کاملParallel Unsupervised Feature Learning with Sparse AutoEncoder
Choosing the right input features for a particular machine learning algorithm is one of the deciding factors for a successful application but often a time consuming manual task. Unsupervised feature learning algorithms provide an alternative solution by automatically learning features; however, they are usually computationally intensive. In this project we explore different implementations of t...
متن کاملUnsupervised Learning aids Prediction: Using Future Representation Learning Variantial Autoencoder for Human Action Prediction
The unsupervised Pretraining method has been widely used in aiding human action recognition. However, existing methods focus on reconstructing the already present frames rather than generating frames which happen in future.In this paper, We propose an improved Variantial Autoencoder model to extract the features with a high connection to the coming scenarios, also known as Predictive Learning. ...
متن کاملAudio Word2Vec: Unsupervised Learning of Audio Segment Representations Using Sequence-to-Sequence Autoencoder
The vector representations of fixed dimensionality for words (in text) offered by Word2Vec have been shown to be very useful in many application scenarios, in particular due to the semantic information they carry. This paper proposes a parallel version, the Audio Word2Vec. It offers the vector representations of fixed dimensionality for variable-length audio segments. These vector representatio...
متن کاملTree-structured Variational Autoencoder
Many kinds of variable-sized data we would like to model contain an internal hierarchical structure in the form of a tree, including source code, formal logical statements, and natural language sentences with parse trees. For such data it is natural to consider a model with matching computational structure. In this work, we introduce a variational autoencoder-based generative model for tree-str...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Proceedings of the ... AAAI Conference on Artificial Intelligence
سال: 2021
ISSN: ['2159-5399', '2374-3468']
DOI: https://doi.org/10.1609/aaai.v35i14.17549