Large language models (LLMs) have achieved remarkable performance on a variety of natural language understanding tasks. However, existing benchmarks are inadequate in measuring the complex logical reasoning capabilities of a model. We present FOLIO, a human-annotated, logically complex and diverse dataset for reasoning in natural language (NL), equipped with first-order logic (FOL) annotations. FOLIO consists of 1,430 examples (unique conclusions), each paired with one of 487 sets of premises used to deductively reason for the validity of each conclusion. The logical correctness of the premises and conclusions is ensured by their FOL annotations, which are automatically verified by an FOL inference engine. In addition to the main NL reasoning task, NL-FOL pairs in FOLIO constitute a new NL-FOL translation dataset. Our experiments on FOLIO systematically evaluate the FOL reasoning ability of supervised fine-tuning on medium-sized language models. For both NL reasoning and NL-FOL translation, we benchmark multiple state-of-the-art language models. Our results show that a subset of FOLIO presents a challenge for one of the most capable {Large Language Model (LLM)} publicly available, GPT-4.
A cornerstone in AI research has been the creation and adoption of standardized training and test datasets to earmark the progress of state-of-the-art models. A particularly successful example is the GLUE dataset for training and evaluating Natural Language Understanding (NLU) models for English. The large body of research around self-supervised BERT-based language models revolved around performance improvements on NLU tasks in GLUE. To evaluate language models in other languages, several language-specific GLUE datasets were created. The area of speech language understanding (SLU) has followed a similar trajectory. The success of large self-supervised models such as wav2vec2 enable creation of speech models with relatively easy to access unlabelled data. These models can then be evaluated on SLU tasks, such as the SUPERB benchmark. In this work, we extend this to Indic languages by releasing the IndicSUPERB benchmark. Specifically, we make the following three contributions. (i) We collect Kathbath containing 1,684 hours of labelled speech data across 12 Indian languages from 1,218 contributors located in 203 districts in India. (ii) Using Kathbath, we create benchmarks across 6 speech tasks: Automatic Speech Recognition, Speaker Verification, Speaker Identification (mono/multi), Language Identification, Query By Example, and Keyword Spotting for 12 languages. (iii) On the released benchmarks, we train and evaluate different self-supervised models alongside a commonly used baseline FBANK. We show that language-specific fine-tuned models are more accurate than baseline on most of the tasks, including a large gap of 76\% for the Language Identification task. However, for speaker identification, self-supervised models trained on large datasets demonstrate an advantage. We hope IndicSUPERB contributes to the progress of developing speech language understanding models for Indian languages.
Transformer-based pretrained language models (LMs) are ubiquitous across natural language understanding, but cannot be applied to long sequences such as stories, scientific articles and long documents, due to their quadratic complexity. While a myriad of efficient transformer variants have been proposed, they are typically based on custom implementations that require expensive pretraining from scratch. In this work, we propose SLED: SLiding-Encoder and Decoder, a simple approach for processing long sequences that re-uses and leverages battle-tested short-text pretrained LMs. Specifically, we partition the input into overlapping chunks, encode each with a short-text LM encoder and use the pretrained decoder to fuse information across chunks (fusion-in-decoder). We illustrate through controlled experiments that SLED offers a viable strategy for long text understanding and evaluate our approach on SCROLLS, a benchmark with seven datasets across a wide range of language understanding tasks. We find that SLED is competitive with specialized models that are up to 50x larger and require a dedicated and expensive pretraining step.
While contextualized word embeddings have been a de-facto standard, learning contextualized phrase embeddings is less explored and being hindered by the lack of a human-annotated benchmark that tests machine understanding of phrase semantics given a context sentence or paragraph (instead of phrases alone). To fill this gap, we propose PiC -- a dataset of ~28K of noun phrases accompanied by their contextual Wikipedia pages and a suite of three tasks for training and evaluating phrase embeddings. Training on PiC improves ranking models' accuracy and remarkably pushes span-selection (SS) models (i.e., predicting the start and end index of the target phrase) near-human accuracy, which is 95% Exact Match (EM) on semantic search given a query phrase and a passage. Interestingly, we find evidence that such impressive performance is because the SS models learn to better capture the common meaning of a phrase regardless of its actual context. SotA models perform poorly in distinguishing two senses of the same phrase in two contexts (~60% EM) and in estimating the similarity between two different phrases in the same context (~70% EM).
We formulate and test a technique to use Emergent Communication (EC) with a pre-trained multilingual model to improve on modern Unsupervised NMT systems, especially for low-resource languages. It has been argued that the current dominant paradigm in NLP of pre-training on text-only corpora will not yield robust natural language understanding systems, and the need for grounded, goal-oriented, and interactive language learning has been high lighted. In our approach, we embed a multilingual model (mBART, Liu et al., 2020) into an EC image-reference game, in which the model is incentivized to use multilingual generations to accomplish a vision-grounded task. The hypothesis is that this will align multiple languages to a shared task space. We present two variants of EC Fine-Tuning (Steinert-Threlkeld et al., 2022), one of which outperforms a backtranslation-only baseline in all four languages investigated, including the low-resource language Nepali.
Automatically classifying electronic health records (EHRs) into diagnostic codes has been challenging to the NLP community. State-of-the-art methods treated this problem as a multilabel classification problem and proposed various architectures to model this problem. However, these systems did not leverage the superb performance of pretrained language models, which achieved superb performance on natural language understanding tasks. Prior work has shown that pretrained language models underperformed on this task with the regular finetuning scheme. Therefore, this paper aims at analyzing the causes of the underperformance and developing a framework for automatic ICD coding with pretrained language models. We spotted three main issues through the experiments: 1) large label space, 2) long input sequences, and 3) domain mismatch between pretraining and fine-tuning. We propose PLMICD, a framework that tackles the challenges with various strategies. The experimental results show that our proposed framework can overcome the challenges and achieves state-of-the-art performance in terms of multiple metrics on the benchmark MIMIC data. The source code is available at https://github.com/MiuLab/PLM-ICD
The probing methodology allows one to obtain a partial representation of linguistic phenomena stored in the inner layers of the neural network, using external classifiers and statistical analysis. Pre-trained transformer-based language models are widely used both for natural language understanding (NLU) and natural language generation (NLG) tasks making them most commonly used for downstream applications. However, little analysis was carried out, whether the models were pre-trained enough or contained knowledge correlated with linguistic theory. We are presenting the chronological probing study of transformer English models such as MultiBERT and T5. We sequentially compare the information about the language learned by the models in the process of training on corpora. The results show that 1) linguistic information is acquired in the early stages of training 2) both language models demonstrate capabilities to capture various features from various levels of language, including morphology, syntax, and even discourse, while they also can inconsistently fail on tasks that are perceived as easy. We also introduce the open-source framework for chronological probing research, compatible with other transformer-based models. https://github.com/EkaterinaVoloshina/chronological_probing
Recently the use of self-attention has yielded to state-of-the-art results in vision-language tasks such as image captioning as well as natural language understanding and generation (NLU and NLG) tasks and computer vision tasks such as image classification. This is since self-attention maps the internal interactions among the elements of input source and target sequences. Although self-attention successfully calculates the attention values and maps the relationships among the elements of input source and target sequence, yet there is no mechanism to control the intensity of attention. In real world, when communicating with each other face to face or vocally, we tend to express different visual and linguistic context with various amounts of intensity. Some words might carry (be spoken with) more stress and weight indicating the importance of that word in the context of the whole sentence. Based on this intuition, we propose Zoneout Dropout Injection Attention Calculation (ZoDIAC) in which the intensities of attention values in the elements of the input sequence are calculated with respect to the context of the elements of input sequence. The results of our experiments reveal that employing ZoDIAC leads to better performance in comparison with the self-attention module in the Transformer model. The ultimate goal is to find out if we could modify self-attention module in the Transformer model with a method that is potentially extensible to other models that leverage on self-attention at their core. Our findings suggest that this particular goal deserves further attention and investigation by the research community. The code for ZoDIAC is available on www.github.com/zanyarz/zodiac .
Large Transformer-based models have exhibited superior performance in various natural language processing and computer vision tasks. However, these models contain enormous amounts of parameters, which restrict their deployment to real-world applications. To reduce the model size, researchers prune these models based on the weights' importance scores. However, such scores are usually estimated on mini-batches during training, which incurs large variability/uncertainty due to mini-batch sampling and complicated training dynamics. As a result, some crucial weights could be pruned by commonly used pruning methods because of such uncertainty, which makes training unstable and hurts generalization. To resolve this issue, we propose PLATON, which captures the uncertainty of importance scores by upper confidence bound (UCB) of importance estimation. In particular, for the weights with low importance scores but high uncertainty, PLATON tends to retain them and explores their capacity. We conduct extensive experiments with several Transformer-based models on natural language understanding, question answering and image classification to validate the effectiveness of PLATON. Results demonstrate that PLATON manifests notable improvement under different sparsity levels. Our code is publicly available at https://github.com/QingruZhang/PLATON.
We propose a method to make natural language understanding models more parameter efficient by storing knowledge in an external knowledge graph (KG) and retrieving from this KG using a dense index. Given (possibly multilingual) downstream task data, e.g., sentences in German, we retrieve entities from the KG and use their multimodal representations to improve downstream task performance. We use the recently released VisualSem KG as our external knowledge repository, which covers a subset of Wikipedia and WordNet entities, and compare a mix of tuple-based and graph-based algorithms to learn entity and relation representations that are grounded on the KG multimodal information. We demonstrate the usefulness of the learned entity representations on two downstream tasks, and show improved performance on the multilingual named entity recognition task by $0.3\%$--$0.7\%$ F1, while we achieve up to $2.5\%$ improvement in accuracy on the visual sense disambiguation task. All our code and data are available in: \url{https://github.com/iacercalixto/visualsem-kg}.
The quest for seeking health information has swamped the web with consumers' health-related questions. Generally, consumers use overly descriptive and peripheral information to express their medical condition or other healthcare needs, contributing to the challenges of natural language understanding. One way to address this challenge is to summarize the questions and distill the key information of the original question. To address this issue, we introduce a new dataset, CHQ-Summ that contains 1507 domain-expert annotated consumer health questions and corresponding summaries. The dataset is derived from the community question-answering forum and therefore provides a valuable resource for understanding consumer health-related posts on social media. We benchmark the dataset on multiple state-of-the-art summarization models to show the effectiveness of the dataset.
In recent years, we witnessed great progress in different tasks of natural language understanding using machine learning. Question answering is one of these tasks which is used by search engines and social media platforms for improved user experience. Arabic is the language of the Holy Qur'an; the sacred text for 1.8 billion people across the world. Arabic is a challenging language for Natural Language Processing (NLP) due to its complex structures. In this article, we describe our attempts at OSACT5 Qur'an QA 2022 Shared Task, which is a question answering challenge on the Holy Qur'an in Arabic. We propose an ensemble learning model based on Arabic variants of BERT models. In addition, we perform post-processing to enhance the model predictions. Our system achieves a Partial Reciprocal Rank (pRR) score of 56.6% on the official test set.
Sequence-to-sequence (seq2seq) learning is a popular fashion for large-scale pretraining language models. However, the prior seq2seq pretraining models generally focus on reconstructive objectives on the decoder side and neglect the effect of encoder-side supervision, which we argue may lead to sub-optimal performance. To verify our hypothesis, we first empirically study the functionalities of the encoder and decoder in seq2seq pretrained language models, and find that the encoder takes an important but under-exploitation role than the decoder regarding the downstream performance and neuron activation. Therefore, we propose an encoding-enhanced seq2seq pretraining strategy, namely E2S2, which improves the seq2seq models via integrating more efficient self-supervised information into the encoders. Specifically, E2S2 adopts two self-supervised objectives on the encoder side from two aspects: 1) locally denoising the corrupted sentence (denoising objective); and 2) globally learning better sentence representations (contrastive objective). With the help of both objectives, the encoder can effectively distinguish the noise tokens and capture high-level (i.e., syntactic and semantic) knowledge, thus strengthening the ability of seq2seq model to accurately achieve the conditional generation. On a large diversity of downstream natural language understanding and generation tasks, E2S2 dominantly improves the performance of its powerful backbone models, e.g., BART and T5. For example, upon BART backbone, we achieve +1.1% averaged gain on the general language understanding evaluation (GLUE) benchmark and +1.75% F_0.5 score improvement on CoNLL2014 dataset. We also provide in-depth analyses to show the improvement stems from better linguistic representation. We hope that our work will foster future self-supervision research on seq2seq language model pretraining.
Instruction tuning is an emergent paradigm in NLP wherein natural language instructions are leveraged with language models to induce zero-shot performance on unseen tasks. Instructions have been shown to enable good performance on unseen tasks and datasets in both large and small language models. Dialogue is an especially interesting area to explore instruction tuning because dialogue systems perform multiple kinds of tasks related to language (e.g., natural language understanding and generation, domain-specific interaction), yet instruction tuning has not been systematically explored for dialogue-related tasks. We introduce InstructDial, an instruction tuning framework for dialogue, which consists of a repository of 48 diverse dialogue tasks in a unified text-to-text format created from 59 openly available dialogue datasets. Next, we explore cross-task generalization ability on models tuned on InstructDial across diverse dialogue tasks. Our analysis reveals that InstructDial enables good zero-shot performance on unseen datasets and tasks such as dialogue evaluation and intent detection, and even better performance in a few-shot setting. To ensure that models adhere to instructions, we introduce novel meta-tasks. We establish benchmark zero-shot and few-shot performance of models trained using the proposed framework on multiple dialogue tasks.
Decision making theories such as Fuzzy-Trace Theory (FTT) suggest that individuals tend to rely on gist, or bottom-line meaning, in the text when making decisions. In this work, we delineate the process of developing GisPy, an open-source tool in Python for measuring the Gist Inference Score (GIS) in text. Evaluation of GisPy on documents in three benchmarks from the news and scientific text domains demonstrates that scores generated by our tool significantly distinguish low vs. high gist documents. Our tool is publicly available to use at: https://github.com/phosseini/GisPy.
Deep Learning has revolutionized the fields of computer vision, natural language understanding, speech recognition, information retrieval and more. Many techniques have evolved over the past decade that made models lighter, faster, and robust with better generalization. However, many deep learning practitioners persist with pre-trained models and architectures trained mostly on standard datasets such as Imagenet, MS-COCO, IMDB-Wiki Dataset, and Kinetics-700 and are either hesitant or unaware of redesigning the architecture from scratch that will lead to better performance. This scenario leads to inefficient models that are not suitable on various devices such as mobile, edge, and fog. In addition, these conventional training methods are of concern as they consume a lot of computing power. In this paper, we revisit various SOTA techniques that deal with architecture efficiency (Global Average Pooling, depth-wise convolutions & squeeze and excitation, Blurpool), learning rate (Cyclical Learning Rate), data augmentation (Mixup, Cutout), label manipulation (label smoothing), weight space manipulation (stochastic weight averaging), and optimizer (sharpness aware minimization). We demonstrate how an efficient deep convolution network can be built in a phased manner by sequentially reducing the number of training parameters and using the techniques mentioned above. We achieved a SOTA accuracy of 99.2% on MNIST data with just 1500 parameters and an accuracy of 86.01% with just over 140K parameters on the CIFAR-10 dataset.
Prompt Tuning has been largely successful as a parameter-efficient method of conditioning large-scale pre-trained language models to perform downstream tasks. Thus far, soft prompt tuning learns a fixed set of task-specific continuous vectors, i.e., soft tokens that remain static across the task samples. A fixed prompt, however, may not generalize well to the diverse kinds of inputs the task comprises. In order to address this, we propose Vector-quantized Input-contextualized Prompts (VIP) as an extension to the soft prompt tuning framework. VIP particularly focuses on two aspects -- contextual prompts that learns input-specific contextualization of the soft prompt tokens through a small-scale sentence encoder and quantized prompts that maps the contextualized prompts to a set of learnable codebook vectors through a Vector quantization network. On various language understanding tasks like SuperGLUE, QA, Relation classification, NER and NLI, VIP outperforms the soft prompt tuning (PT) baseline by an average margin of 1.19%. Further, our generalization studies show that VIP learns more robust prompt representations, surpassing PT by a margin of 0.6% - 5.3% on Out-of-domain QA and NLI tasks respectively, and by 0.75% on Multi-Task setup over 4 tasks spanning across 12 domains.
Recent model pruning methods have demonstrated the ability to remove redundant parameters without sacrificing model performance. Common methods remove redundant parameters according to the parameter sensitivity, a gradient-based measure reflecting the contribution of the parameters. In this paper, however, we argue that redundant parameters can be trained to make beneficial contributions. We first highlight the large sensitivity (contribution) gap among high-sensitivity and low-sensitivity parameters and show that the model generalization performance can be significantly improved after balancing the contribution of all parameters. Our goal is to balance the sensitivity of all parameters and encourage all of them to contribute equally. We propose a general task-agnostic method, namely intra-distillation, appended to the regular training loss to balance parameter sensitivity. Moreover, we also design a novel adaptive learning method to control the strength of intra-distillation loss for faster convergence. Our experiments show the strong effectiveness of our methods on machine translation, natural language understanding, and zero-shot cross-lingual transfer across up to 48 languages, e.g., a gain of 3.54 BLEU on average across 8 language pairs from the IWSLT'14 translation dataset.
Transformers, currently the state-of-the-art in natural language understanding (NLU) tasks, are prone to generate uncalibrated predictions or extreme probabilities, making the process of taking different decisions based on their output relatively difficult. In this paper we propose to build several inductive Venn--ABERS predictors (IVAP), which are guaranteed to be well calibrated under minimal assumptions, based on a selection of pre-trained transformers. We test their performance over a set of diverse NLU tasks and show that they are capable of producing well-calibrated probabilistic predictions that are uniformly spread over the [0,1] interval -- all while retaining the original model's predictive accuracy.
Solving crossword puzzles requires diverse reasoning capabilities, access to a vast amount of knowledge about language and the world, and the ability to satisfy the constraints imposed by the structure of the puzzle. In this work, we introduce solving crossword puzzles as a new natural language understanding task. We release the specification of a corpus of crossword puzzles collected from the New York Times daily crossword spanning 25 years and comprised of a total of around nine thousand puzzles. These puzzles include a diverse set of clues: historic, factual, word meaning, synonyms/antonyms, fill-in-the-blank, abbreviations, prefixes/suffixes, wordplay, and cross-lingual, as well as clues that depend on the answers to other clues. We separately release the clue-answer pairs from these puzzles as an open-domain question answering dataset containing over half a million unique clue-answer pairs. For the question answering task, our baselines include several sequence-to-sequence and retrieval-based generative models. We also introduce a non-parametric constraint satisfaction baseline for solving the entire crossword puzzle. Finally, we propose an evaluation framework which consists of several complementary performance metrics.