Abstract

Topic diversion occurs frequently with engaging open-domain dialogue systems like virtual assistants. The balance between staying on topic and rectifying the topic drift is important for a good collaborative system. In this paper, we present a model which uses a finetuned XLNet-base to classify the utterances pertaining to the major topic of conversation and those which are not, with a precision of 84%. We propose a preliminary study, classifying utterances into major, minor and offtopics, which further extends into a system initiative for diversion rectification. A case study was conducted where a system initiative is emulated as a response to the user going off-topic, mimicking a common occurrence of mixed initiative present in natural human-human conversation. This task of classifying utterances into those which belong to the major theme or not, would also help us in identification of relevant sentences for tasks like dialogue summarization and information extraction from conversations.

Abstract

This paper describes SimpleNER, a model developed for the sentence simplification task at GEM-2021. Our system is a monolingual Seq2Seq Transformer architecture that uses control tokens pre-pended to the data, allowing the model to shape the generated simplifications according to user desired attributes. Additionally, we show that NER-tagging the training data before use helps stabilize the effect of the control tokens and significantly improves the overall performance of the system. We also employ pretrained embeddings to reduce data sparsity and allow the model to produce more generalizable outputs.

Abstract

Aspect extraction is not a well-explored topic in Hindi, with only one corpus having been developed for the task. In this paper, we discuss the merits of the existing corpus in terms of quality, size, sparsity, and performance in aspect extraction tasks using established models. To provide a better baseline corpus for aspect extraction, we translate the SemEval 2014 aspect-based sentiment analysis dataset and annotate the aspects in that data. We provide rigorous guidelines and a replicable methodology for this task. We quantitatively evaluate the translations and annotations using inter-annotator agreement scores. We also evaluate our dataset using state-of-the-art neural aspect extraction models in both monolingual and multilingual settings and show that the models perform far better on our corpus than on the existing Hindi dataset. With this, we establish our corpus as the gold-standard aspect extraction dataset in Hindi.

Abstract

Multiple parallel attention mechanisms that use multiple attention heads facilitate greater performance of the Transformer model for various applications e.g., Neural Machine Translation (NMT), text classification. In multi-head attention mechanism, different heads attend to different parts of the input. However, the limitation is that multiple heads might attend to the same part of the input, resulting in multiple heads being redundant. Thus, the model resources are under-utilized. One approach to avoid this is to prune least important heads based on certain importance score. In this work, we focus on designing a Dynamic Head Importance Computation Mechanism (DHICM) to dynamically calculate the importance of a head with respect to the input. Our insight is to design an additional attention layer together with multi-head attention, and utilize the outputs of the multi-head attention along with the input, to compute the importance for each head. Additionally, we add an extra loss function to prevent the model from assigning same score to all heads, to identify more important heads and improvise performance. We analyzed performance of DHICM for NMT with different languages. Experiments on different datasets show that DHICM outperforms traditional Transformer-based approach by large margin, especially, when less training data is available.

Abstract

The extensive rise in consumption of online social media (OSMs) by a large number of people poses a critical problem of curbing the spread of hateful content on these platforms. With the growing usage of OSMs in multiple languages, the task of detecting and characterizing hate becomes more complex. The subtle variations of code-mixed texts along with switching scripts only add to the complexity. This paper presents a solution for the HASOC 2021 Multilingual Twitter Hate-Speech Detection challenge by team PreCog IIIT Hyderabad. We adopt a multilingual transformer based approach and describe our architecture for all 6 sub-tasks as part of the challenge. Out of the 6 teams that participated in all the sub tasks, our submissions rank 3rd overall.

Abstract

Social media platforms have democratized the publication process resulting into easy and viral propagation of information. Oftentimes this misinformation is accompanied by misleading or doctored images that quickly circulate across the internet and reach many unsuspecting users. Several manual as well as automated efforts have been undertaken in the past to solve this critical problem. While manual efforts cannot keep up with the rate at which this content is churned out, many automated approaches only leverage concatenation (of the image and text representations) thereby failing to build effective crossmodal embeddings. Architectures like this fail in many cases because the text or image doesn’t need to be false for the corresponding text, image pair to be misinformation. While some recent work attempts to use attention techniques to compute a crossmodal representation using pretrained text and image embeddings, we show a more effective approach towards utilizing such pretrained embeddings to build richer representations that can be classified better. This involves several challenges like how to handle text variations on Twitter and Weibo, how to encode the image information and how to leverage the text and image encodings together effectively. Our architecture, SCATE (Shared Cross Attention Transformer Encoders), leverages deep convolutional neural networks and transformer-based methods to encode image and text information utilizing crossmodal attention and shared layers for the two modalities. Our experiments with three popular benchmark datasets (Twitter, WeiboA and WeiboB) show that our proposed methods outperform the state-of-the-art meth

Abstract

Early detection of cardiovascular diseases is crucial for effective treatment and an electrocardiogram (ECG) is pivotal for diagnosis. The accuracy of Deep Learning based methods for ECG signal classification has progressed in recent years to reach cardiologist-level performance. In clinical settings, a cardiologist makes a diagnosis based on the standard 12-channel ECG recording. Automatic analysis of ECG recordings from a multiple-channel perspective has not been given enough attention, so it is essential to analyze an ECG recording from a multiple-channel perspective. We propose a model that leverages the multiple-channel information available in the standard 12-channel ECG recordings and learns patterns at the beat, rhythm, and channel level. The experimental results show that our model achieved a macro-averaged ROC-AUC score of 0.9216, mean accuracy of 88.85% and a maximum F1 score of 0.8057 on the PTB-XL dataset. The attention visualization results from the interpretable model are compared against the cardiologist’s guidelines to validate the correctness and usability

Abstract

Long standing challenges in adaptive bipedal walking control (i.e. control taking care of unknown robot parameters) were to unify the control design instead of designing multiple controllers for different walking phases as well as to bypass computing constraint forces, since it often leads to complex designs. A few attempts to design a single controller for all walking phases ignored or oversimplified the constraint forces. However, these forces are state-dependent and may lead to conservative performance or instability if not countered properly. This work proposes an innovative adaptive control method, based on artificial time delay control, which covers the entire bipedal walking phase and provides robustness against state-dependent unmodelled dynamics such as constraint forces and external impulsive forces arising during walking. Studies using a high fidelity simulator under various forms of disturbances show the effectiveness of the proposed design over the state of the art.

Abstract

A crucial challenge in maintaining formation in an autonomous vehicle platoon rests on designing a suitable control scheme that can tackle external disturbances and uncertain system parameters, especially the friction forces between wheel and ground, which vary with change in road surface, wear in tires and speed of the vehicle. State-of-the-art adaptive controller negotiating such uncertainties can typically handle a priori bounded (or non state-dependent) uncertainties. However, besides the difficulty in accurate modelling and identification of frictional forces, in general, these forces are state-dependent and cannot be a priori bounded (e.g., viscous friction). This paper proposes an adaptive sliding mode controller for nonholonomic wheeled mobile robot based vehicle platoon which can handle unknown complex behaviour of frictional forces without a priori knowledge of its parameters and structures. The effectiveness of the proposed controller is verified in Gazebo simulation in comparison with the state of the art.

Abstract

A major challenge that deep learning systems face is the Catastrophic Forgetting (CF) phenomenon that is observed when finetuning is used to try and adapt a system to a new task or a sequence of datasets with different distributions. CF refers to the significant degradation in performance on the old task/dataset. In this paper, a novel approach is proposed to address CF in computer aided diagnosis (CAD) system design in the medical domain. CAD systems often need to handle a sequence of datasets collected over time from different sites with different imaging parameters/populations. The solution we propose is to move samples from all the datasets closer to a common manifold via a reformer at the front end of a CAD system. The utility of this approach is demonstrated on two common tasks, namely segmentation and classification, using publicly available datasets. Results of extensive experiments show that manifold learning can yield about 74% improvement on an average in the reduction of CF over the baseline fine-tuning process and the state-of-the-art regularization based methods. The results also indicate that a Reformer when used in conjunction with the state-of-the-art regularization methods, has the potential to yield further improvement in CF reduction.