More Like this

1. Unsupervised Parsing with S-DIORA: Single Tree Encoding for Deep Inside-Outside Recursive Autoencoders

   https://doi.org/10.18653/v1/2020.emnlp-main.392  

   Drozdov, Andrew; Rongali, Subendhu; Chen, Yi-Pei; O’Gorman, Tim; Iyyer, Mohit; McCallum, Andrew (November 2020, Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP))

   null (Ed.)

   The deep inside-outside recursive autoencoder (DIORA; Drozdov et al. 2019) is a self-supervised neural model that learns to induce syntactic tree structures for input sentences *without access to labeled training data*. In this paper, we discover that while DIORA exhaustively encodes all possible binary trees of a sentence with a soft dynamic program, its vector averaging approach is locally greedy and cannot recover from errors when computing the highest scoring parse tree in bottom-up chart parsing. To fix this issue, we introduce S-DIORA, an improved variant of DIORA that encodes a single tree rather than a softly-weighted mixture of trees by employing a hard argmax operation and a beam at each cell in the chart. Our experiments show that through *fine-tuning* a pre-trained DIORA with our new algorithm, we improve the state of the art in *unsupervised* constituency parsing on the English WSJ Penn Treebank by 2.2-6% F1, depending on the data used for fine-tuning. 

   more » « less
2. The lottery ticket hypothesis for pre-trained BERT networks

   Chen, Tianlong; Frankle, Jonathan; Chang, Shiyu; Liu, Sijia; Zhang, Yang; Wang, Zhangyang; Carbin, Michael (December 2020, Advances in neural information processing systems)

   null (Ed.)

   In natural language processing (NLP), enormous pre-trained models like BERT have become the standard starting point for training on a range of downstream tasks, and similar trends are emerging in other areas of deep learning. In parallel, work on the lottery ticket hypothesis has shown that models for NLP and computer vision contain smaller matching subnetworks capable of training in isolation to full accuracy and transferring to other tasks. In this work, we combine these observations to assess whether such trainable, transferrable subnetworks exist in pre-trained BERT models. For a range of downstream tasks, we indeed find matching subnetworks at 40% to 90% sparsity. We find these subnetworks at (pre-trained) initialization, a deviation from prior NLP research where they emerge only after some amount of training. Subnetworks found on the masked language modeling task (the same task used to pre-train the model) transfer universally; those found on other tasks transfer in a limited fashion if at all. As large-scale pre-training becomes an increasingly central paradigm in deep learning, our results demonstrate that the main lottery ticket observations remain relevant in this context. 

   more » « less
3. Surface Statistics of an Unknown Language Indicate How to Parse It

   https://doi.org/10.1162/tacl_a_00248  

   Wang, Dingquan; Eisner, Jason (December 2018, Transactions of the Association for Computational Linguistics)

   We introduce a novel framework for delexicalized dependency parsing in a new language. We show that useful features of the target language can be extracted automatically from an unparsed corpus, which consists only of gold part-of-speech (POS) sequences. Providing these features to our neural parser enables it to parse sequences like those in the corpus. Strikingly, our system has no supervision in the target language. Rather, it is a multilingual system that is trained end-to-end on a variety of other languages, so it learns a feature extractor that works well. We show experimentally across multiple languages: (1) Features computed from the unparsed corpus improve parsing accuracy. (2) Including thousands of synthetic languages in the training yields further improvement. (3) Despite being computed from unparsed corpora, our learned task-specific features beat previous work’s interpretable typological features that require parsed corpora or expert categorization of the language. Our best method improved attachment scores on held-out test languages by an average of 5.6 percentage points over past work that does not inspect the unparsed data (McDonald et al., 2011), and by 20.7 points over past “grammar induction” work that does not use training languages (Naseem et al., 2010). 

   more » « less
4. Bidirectional Language Models Are Also Few-shot Learners

   Patel, Ajay; Li, Bryan; Rasooli, Mohammad Sadegh; Constant, Noah; Raffel, Colin; Callison-Burch, Chris (May 2023, The Eleventh International Conference on Learning Representations (ICLR 2023))

   Large language models such as GPT-3 (Brown et al., 2020) can perform arbitrary tasks without undergoing fine-tuning after being prompted with only a few labeled examples. An arbitrary task can be reformulated as a natural language prompt, and a language model can be asked to generate the completion, indirectly performing the task in a paradigm known as prompt-based learning. To date, emergent prompt-based learning capabilities have mainly been demonstrated for unidirectional language models. However, bidirectional language models pre-trained on denoising objectives such as masked language modeling produce stronger learned representations for transfer learning. This motivates the possibility of prompting bidirectional models, but their pre-training objectives have made them largely incompatible with the existing prompting paradigm. We present SAP (Sequential Autoregressive Prompting), a technique that enables the prompting of bidirectional models. Utilizing the machine translation task as a case study, we prompt the bidirectional mT5 model (Xue et al., 2021) with SAP and demonstrate its few-shot and zero-shot translations outperform the few-shot translations of unidirectional models like GPT-3 and XGLM (Lin et al., 2021), despite mT5's approximately 50% fewer parameters. We further show SAP is effective on question answering and summarization. For the first time, our results demonstrate prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models. 

   more » « less
5. Fast and Space-Efficient Construction of AVL Grammars from the LZ77 Parsing.

   https://doi.org/10.4230/LIPIcs.ESA.2021.56  

   Kempa, Dominik and (January 2021, 29th Annual European Symposium on Algorithms)

   Mutzel, Petra and (Ed.)

   Grammar compression is, next to Lempel-Ziv (LZ77) and run-length Burrows-Wheeler transform (RLBWT), one of the most flexible approaches to representing and processing highly compressible strings. The main idea is to represent a text as a context-free grammar whose language is precisely the input string. This is called a straight-line grammar (SLG). An AVL grammar, proposed by Rytter [Theor. Comput. Sci., 2003] is a type of SLG that additionally satisfies the AVL property: the heights of parse trees for children of every nonterminal differ by at most one. In contrast to other SLG constructions, AVL grammars can be constructed from the LZ77 parsing in compressed time: 𝒪(z log n) where z is the size of the LZ77 parsing and n is the length of the input text. Despite these advantages, AVL grammars are thought to be too large to be practical. We present a new technique for rapidly constructing a small AVL grammar from an LZ77 or LZ77-like parse. Our algorithm produces grammars that are always at least five times smaller than those produced by the original algorithm, and usually not more than double the size of grammars produced by the practical Re-Pair compressor [Larsson and Moffat, Proc. IEEE, 2000]. Our algorithm also achieves low peak RAM usage. By combining this algorithm with recent advances in approximating the LZ77 parsing, we show that our method has the potential to construct a run-length BWT in about one third of the time and peak RAM required by other approaches. Overall, we show that AVL grammars are surprisingly practical, opening the door to much faster construction of key compressed data structures 

   more » « less