Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG-bench). BIG-bench currently consists of 204 tasks, contributed by 442 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood development, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting.

One to rule them all: Towards Joint Indic Language Hate Speech Detection

This paper is a contribution to the Hate Speech and Offensive Content Identification in Indo-European Languages (HASOC) 2021 shared task. Social media today is a hotbed of toxic and hateful conversations, in various languages. Recent news reports have shown that current models struggle to automatically identify hate posted in minority languages. Therefore, efficiently curbing hate speech is a critical challenge and problem of interest. We present a multilingual architecture using state-of-the-art transformer language models to jointly learn hate and offensive speech detection across three languages namely, English, Hindi, and Marathi. On the provided testing corpora, we achieve Macro F1 scores of 0.7996, 0.7748, 0.8651 for sub-task 1A and 0.6268, 0.5603 during the fine-grained classification of sub-task 1B. These results show the efficacy of exploiting a multilingual training scheme.

Translational Equivariance in Kernelizable Attention

While Transformer architectures have show remarkable success, they are bound to the computation of all pairwise interactions of input element and thus suffer from limited scalability. Recent work has been successful by avoiding the computation of the complete attention matrix, yet leads to problems down the line. The absence of an explicit attention matrix makes the inclusion of inductive biases relying on relative interactions between elements more challenging. An extremely powerful inductive bias is translational equivariance, which has been conjectured to be responsible for much of the success of Convolutional Neural Networks on image recognition tasks. In this work we show how translational equivariance can be implemented in efficient Transformers based on kernelizable attention - Performers. Our experiments highlight that the devised approach significantly improves robustness of Performers to shifts of input images compared to their naive application. This represents an important step on the path of replacing Convolutional Neural Networks with more expressive Transformer architectures and will help to improve sample efficiency and robustness in this realm.

Indic-Transformers: An Analysis of Transformer Language Models for Indian Languages

Language models based on the Transformer architecture have achieved state-of-the-art performance on a wide range of NLP tasks such as text classification, question-answering, and token classification. However, this performance is usually tested and reported on high-resource languages, like English, French, Spanish, and German. Indian languages, on the other hand, are underrepresented in such benchmarks. Despite some Indian languages being included in training multilingual Transformer models, they have not been the primary focus of such work. In order to evaluate the performance on Indian languages specifically, we analyze these language models through extensive experiments on multiple downstream tasks in Hindi, Bengali, and Telugu language. Here, we compare the efficacy of fine-tuning model parameters of pre-trained models against that of training a language model from scratch. Moreover, we empirically argue against the strict dependency between the dataset size and model performance, but rather encourage task-specific model and method selection. We achieve state-of-the-art performance on Hindi and Bengali languages for text classification task. Finally, we present effective strategies for handling the modeling of Indian languages and we release our model checkpoints for the community : https://huggingface.co/neuralspace-reverie.

End to End Binarized Neural Networks for Text Classification

Deep neural networks have demonstrated their superior performance in almost every Natural Language Processing task, however, their increasing complexity raises concerns. In particular, these networks require high expenses on computational hardware, and training budget is a concern for many. Even for a trained network, the inference phase can be too demanding for resource-constrained devices, thus limiting its applicability. The state-of-the-art transformer models are a vivid example. Simplifying the computations performed by a network is one way of relaxing the complexity requirements. In this paper, we propose an end to end binarized neural network architecture for the intent classification task. In order to fully utilize the potential of end to end binarization, both input representations (vector embeddings of tokens statistics) and the classifier are binarized. We demonstrate the efficiency of such architecture on the intent classification of short texts over three datasets and for text classification with a larger dataset. The proposed architecture achieves comparable to the state-of-the-art results on standard intent classification datasets while utilizing ~ 20-40% lesser memory and training time. Furthermore, the individual components of the architecture, such as binarized vector embeddings of documents or binarized classifiers, can be used separately with not necessarily fully binary architectures.

HyperEmbed: Tradeoffs Between Resources and Performance in NLP Tasks with Hyperdimensional Computing enabled Embedding of n-gram Statistics

Recent advances in Deep Learning have led to a significant performance increase on several NLP tasks, however, the models become more and more computationally demanding. Therefore, this paper tackles the domain of computationally efficient algorithms for NLP tasks. In particular, it investigates distributed representations of n-gram statistics of texts. The representations are formed using hyperdimensional computing enabled embedding. These representations then serve as features, which are used as input to standard classifiers. We investigate the applicability of the embedding on one large and three small standard datasets for classification tasks using nine classifiers. The embedding achieved on par F1 scores while decreasing the time and memory requirements by several times compared to the conventional n-gram statistics, e.g., for one of the classifiers on a small dataset, the memory reduction was 6.18 times; while train and test speed-ups were 4.62 and 3.84 times, respectively. For many classifiers on the large dataset, the memory reduction was about 100 times and train and test speed-ups were over 100 times. More importantly, the usage of distributed representations formed via hyperdimensional computing allows dissecting the strict dependency between the dimensionality of the representation and the parameters of n-gram statistics, thus, opening a room for tradeoffs.

ProbAct: A Probabilistic Activation Function for Deep Neural Networks

Activation functions play an important role in the training of artificial neural networks and the Rectified Linear Unit (ReLU) has been the mainstream in recent years. Most of the activation functions currently used are deterministic in nature, whose input-output relationship is fixed. In this work, we propose a probabilistic activation function, called ProbAct. The output value of ProbAct is sampled from a normal distribution, with the mean value same as the output of ReLU and with a fixed or trainable variance for each element. In the trainable ProbAct, the variance of the activation distribution is trained through back-propagation. We also show that the stochastic perturbation through ProbAct is a viable generalization technique that can prevent overfitting. In our experiments, we demonstrate that when using ProbAct, it is possible to boost the image classification performance on CIFAR-10, CIFAR-100, and STL-10 datasets.

A Comprehensive guide to Bayesian Convolutional Neural Network with Variational Inference

Artificial Neural Networks are connectionist systems that perform a given task by learning on examples without having prior knowledge about the task. This is done by finding an optimal point estimate for the weights in every node. Generally, the network using point estimates as weights perform well with large datasets, but they fail to express uncertainty in regions with little or no data, leading to overconfident decisions. In this paper, Bayesian Convolutional Neural Network (BayesCNN) using Variational Inference is proposed, that introduces probability distribution over the weights. Furthermore, the proposed BayesCNN architecture is applied to tasks like Image Classification, Image Super-Resolution and Generative Adversarial Networks. The results are compared to point-estimates based architectures on MNIST, CIFAR-10 and CIFAR-100 datasets for Image CLassification task, on BSD300 dataset for Image Super Resolution task and on CIFAR10 dataset again for Generative Adversarial Network task. BayesCNN is based on Bayes by Backprop which derives a variational approximation to the true posterior. We, therefore, introduce the idea of applying two convolutional operations, one for the mean and one for the variance. Our proposed method not only achieves performances equivalent to frequentist inference in identical architectures but also incorporate a measurement for uncertainties and regularisation. It further eliminates the use of dropout in the model. Moreover, we predict how certain the model prediction is based on the epistemic and aleatoric uncertainties and empirically show how the uncertainty can decrease, allowing the decisions made by the network to become more deterministic as the training accuracy increases. Finally, we propose ways to prune the Bayesian architecture and to make it more computational and time effective.

Subword Semantic Hashing for Intent Classification on Small Datasets

In this paper, we introduce the use of Semantic Hashing as embedding for the task of Intent Classification and outperform previous state-of-the-art methods on three frequently used benchmarks. Intent Classification on a small dataset is a challenging task for data-hungry state-of-the-art Deep Learning based systems. Semantic Hashing is an attempt to overcome such a challenge and learn robust text classification. Current word embedding based methods are dependent on vocabularies. One of the major drawbacks of such methods is out-of-vocabulary terms, especially when having small training datasets and using a wider vocabulary. This is the case in Intent Classification for chatbots, where typically small datasets are extracted from internet communication. Two problems arise by the use of internet communication. First, such datasets miss a lot of terms in the vocabulary to use word embeddings efficiently. Second, users frequently make spelling errors. Typically, the models for intent classification are not trained with spelling errors and it is difficult to think about ways in which users will make mistakes. Models depending on a word vocabulary will always face such issues. An ideal classifier should handle spelling errors inherently. With Semantic Hashing, we overcome these challenges and achieve state-of-the-art results on three datasets: AskUbuntu, Chatbot, and Web Application. Our benchmarks are available online: https://github.com/kumar-shridhar/Know-Your-Intent

Bayesian Convolutional Neural Networks

We introduce Bayesian Convolutional Neural Networks (BayesCNNs), a variant of Convolutional Neural Networks (CNNs) which is built upon Bayes by Backprop. We demonstrate how this novel reliable variational inference method can serve as a fundamental construct for various network architectures. On multiple datasets in supervised learning settings (MNIST, CIFAR-10, CIFAR-100, and STL-10), our proposed variational inference method achieves performances equivalent to frequentist inference in identical architectures, while a measurement for uncertainties and a regularisation are incorporated naturally. In the past, Bayes by Backprop has been successfully implemented in feedforward and recurrent neural networks, but not in convolutional ones. This work symbolises the extension of Bayesian neural networks which encompasses all three aforementioned types of network architectures now.