Sri Harsha Dumpala

Sri Harsha Dumpala, Chandramouli Sastry, Sageev Oore

In this paper, we study the application of Test-Time Training (TTT) as a solution to handling distribution shifts in speech applications. In particular, we introduce distribution-shifts to the test datasets of standard speech-classification tasks -- for example, speaker-identification and emotion-detection -- and explore how Test-Time Training (TTT) can help adjust to the distribution-shift. In our experiments that include distribution shifts due to background noise and natural variations in speech such as gender and age, we identify some key-challenges with TTT including sensitivity to optimization hyperparameters (e.g., number of optimization steps and subset of parameters chosen for TTT) and scalability (e.g., as each example gets its own set of parameters, TTT is not scalable). Finally, we propose using BitFit -- a parameter-efficient fine-tuning algorithm proposed for text applications that only considers the bias parameters for fine-tuning -- as a solution to the aforementioned challenges and demonstrate that it is consistently more stable than fine-tuning all the parameters of the model.

Chandramouli Sastry, Sri Harsha Dumpala, Sageev Oore

We introduce DiffAug, a simple and efficient diffusion-based augmentation technique to train image classifiers for the crucial yet challenging goal of improved classifier robustness. Applying DiffAug to a given example consists of one forward-diffusion step followed by one reverse-diffusion step. Using both ResNet-50 and Vision Transformer architectures, we comprehensively evaluate classifiers trained with DiffAug and demonstrate the surprising effectiveness of single-step reverse diffusion in improving robustness to covariate shifts, certified adversarial accuracy and out of distribution detection. When we combine DiffAug with other augmentations such as AugMix and DeepAugment we demonstrate further improved robustness. Finally, building on this approach, we also improve classifier-guided diffusion wherein we observe improvements in: (i) classifier-generalization, (ii) gradient quality (i.e., improved perceptual alignment) and (iii) image generation performance. We thus introduce a computationally efficient technique for training with improved robustness that does not require any additional data, and effectively complements existing augmentation approaches.

Sri Harsha Dumpala, Aman Jaiswal, Chandramouli Sastry et al.

Despite their remarkable successes, state-of-the-art language models face challenges in grasping certain important semantic details. This paper introduces the VISLA (Variance and Invariance to Semantic and Lexical Alterations) benchmark, designed to evaluate the semantic and lexical understanding of language models. VISLA presents a 3-way semantic (in)equivalence task with a triplet of sentences associated with an image, to evaluate both vision-language models (VLMs) and unimodal language models (ULMs). An evaluation involving 34 VLMs and 20 ULMs reveals surprising difficulties in distinguishing between lexical and semantic variations. Spatial semantics encoded by language models also appear to be highly sensitive to lexical information. Notably, text encoders of VLMs demonstrate greater sensitivity to semantic and lexical variations than unimodal text encoders. Our contributions include the unification of image-to-text and text-to-text retrieval tasks, an off-the-shelf evaluation without fine-tuning, and assessing LMs' semantic (in)variance in the presence of lexical alterations. The results highlight strengths and weaknesses across diverse vision and unimodal language models, contributing to a deeper understanding of their capabilities. % VISLA enables a rigorous evaluation, shedding light on language models' capabilities in handling semantic and lexical nuances. Data and code will be made available at https://github.com/Sri-Harsha/visla_benchmark.

Sri Harsha Dumpala, Aman Jaiswal, Chandramouli Sastry et al.

Despite the significant influx of prompt-tuning techniques for generative vision-language models (VLMs), it remains unclear how sensitive these models are to lexical and semantic alterations in prompts. In this paper, we evaluate the ability of generative VLMs to understand lexical and semantic changes in text using the SugarCrepe++ dataset. We analyze the sensitivity of VLMs to lexical alterations in prompts without corresponding semantic changes. Our findings demonstrate that generative VLMs are highly sensitive to such alterations. Additionally, we show that this vulnerability affects the performance of techniques aimed at achieving consistency in their outputs.

Sri Harsha Dumpala, David Arps, Sageev Oore et al.

Vision-language models (VLMs), serve as foundation models for multi-modal applications such as image captioning and text-to-image generation. Recent studies have highlighted limitations in VLM text encoders, particularly in areas like compositionality and semantic understanding, though the underlying reasons for these limitations remain unclear. In this work, we aim to address this gap by analyzing the syntactic information, one of the fundamental linguistic properties, encoded by the text encoders of VLMs. We perform a thorough analysis comparing VLMs with different objective functions, parameter size and training data size, and with uni-modal language models (ULMs) in their ability to encode syntactic knowledge. Our findings suggest that ULM text encoders acquire syntactic information more effectively than those in VLMs. The syntactic information learned by VLM text encoders is shaped primarily by the pre-training objective, which plays a more crucial role than other factors such as model architecture, model size, or the volume of pre-training data. Models exhibit different layer-wise trends where CLIP performance dropped across layers while for other models, middle layers are rich in encoding syntactic knowledge.

Sri Harsha Dumpala, Chandramouli Shama Sastry, Rudolf Uher et al.

Previous works on depression detection use datasets collected in similar environments to train and test the models. In practice, however, the train and test distributions cannot be guaranteed to be identical. Distribution shifts can be introduced due to variations such as recording environment (e.g., background noise) and demographics (e.g., gender, age, etc). Such distributional shifts can surprisingly lead to severe performance degradation of the depression detection models. In this paper, we analyze the application of test-time training (TTT) to improve robustness of models trained for depression detection. When compared to regular testing of the models, we find TTT can significantly improve the robustness of the model under a variety of distributional shifts introduced due to: (a) background-noise, (b) gender-bias, and (c) data collection and curation procedure (i.e., train and test samples are from separate datasets).

Sri Harsha Dumpala, Katerina Dikaios, Abraham Nunes et al.

Depression, a prevalent mental health disorder impacting millions globally, demands reliable assessment systems. Unlike previous studies that focus solely on either detecting depression or predicting its severity, our work identifies individual symptoms of depression while also predicting its severity using speech input. We leverage self-supervised learning (SSL)-based speech models to better utilize the small-sized datasets that are frequently encountered in this task. Our study demonstrates notable performance improvements by utilizing SSL embeddings compared to conventional speech features. We compare various types of SSL pretrained models to elucidate the type of speech information (semantic, speaker, or prosodic) that contributes the most in identifying different symptoms. Additionally, we evaluate the impact of combining multiple SSL embeddings on performance. Furthermore, we show the significance of multi-task learning for identifying depressive symptoms effectively.

Sebastian Rodriguez, Sri Harsha Dumpala, Katerina Dikaios et al.

Current automatic depression detection systems provide predictions directly without relying on the individual symptoms/items of depression as denoted in the clinical depression rating scales. In contrast, clinicians assess each item in the depression rating scale in a clinical setting, thus implicitly providing a more detailed rationale for a depression diagnosis. In this work, we make a first step towards using the acoustic features of speech to predict individual items of the depression rating scale before obtaining the final depression prediction. For this, we use convolutional (CNN) and recurrent (long short-term memory (LSTM)) neural networks. We consider different approaches to learning the temporal context of speech. Further, we analyze two variants of voting schemes for individual item prediction and depression detection. We also include an animated visualization that shows an example of item prediction over time as the speech progresses.

Sri Harsha Dumpala, Aman Jaiswal, Chandramouli Sastry et al.

Despite their remarkable successes, state-of-the-art large language models (LLMs), including vision-and-language models (VLMs) and unimodal language models (ULMs), fail to understand precise semantics. For example, semantically equivalent sentences expressed using different lexical compositions elicit diverging representations. The degree of this divergence and its impact on encoded semantics is not very well understood. In this paper, we introduce the SUGARCREPE++ dataset to analyze the sensitivity of VLMs and ULMs to lexical and semantic alterations. Each sample in SUGARCREPE++ dataset consists of an image and a corresponding triplet of captions: a pair of semantically equivalent but lexically different positive captions and one hard negative caption. This poses a 3-way semantic (in)equivalence problem to the language models. We comprehensively evaluate VLMs and ULMs that differ in architecture, pre-training objectives and datasets to benchmark the performance of SUGARCREPE++ dataset. Experimental results highlight the difficulties of VLMs in distinguishing between lexical and semantic variations, particularly in object attributes and spatial relations. Although VLMs with larger pre-training datasets, model sizes, and multiple pre-training objectives achieve better performance on SUGARCREPE++, there is a significant opportunity for improvement. We show that all the models which achieve better performance on compositionality datasets need not perform equally well on SUGARCREPE++, signifying that compositionality alone may not be sufficient for understanding semantic and lexical alterations. Given the importance of the property that the SUGARCREPE++ dataset targets, it serves as a new challenge to the vision-and-language community.

Jason d'Eon, Sri Harsha Dumpala, Chandramouli Shama Sastry et al.

In this paper, we propose a new compositional tool that will generate a musical outline of speech recorded/provided by the user for use as a musical building block in their compositions. The tool allows any user to use their own speech to generate musical material, while still being able to hear the direct connection between their recorded speech and the resulting music. The tool is built on our proposed pipeline. This pipeline begins with speech-based signal processing, after which some simple musical heuristics are applied, and finally these pre-processed signals are passed through Transformer models trained on new musical tasks. We illustrate the effectiveness of our pipeline -- which does not require a paired dataset for training -- through examples of music created by musicians making use of our tool.

Sri Harsha Dumpala, Sebastian Rodriguez, Sheri Rempel et al.

Depression detection from speech has attracted a lot of attention in recent years. However, the significance of speaker-specific information in depression detection has not yet been explored. In this work, we analyze the significance of speaker embeddings for the task of depression detection from speech. Experimental results show that the speaker embeddings provide important cues to achieve state-of-the-art performance in depression detection. We also show that combining conventional OpenSMILE and COVAREP features, which carry complementary information, with speaker embeddings further improves the depression detection performance. The significance of temporal context in the training of deep learning models for depression detection is also analyzed in this paper.

Sri Harsha Dumpala, Imran Sheikh, Rupayan Chakraborty et al.

Naturally introduced perturbations in audio signal, caused by emotional and physical states of the speaker, can significantly degrade the performance of Automatic Speech Recognition (ASR) systems. In this paper, we propose a front-end based on Cycle-Consistent Generative Adversarial Network (CycleGAN) which transforms naturally perturbed speech into normal speech, and hence improves the robustness of an ASR system. The CycleGAN model is trained on non-parallel examples of perturbed and normal speech. Experiments on spontaneous laughter-speech and creaky-speech datasets show that the performance of four different ASR systems improve by using speech obtained from CycleGAN based front-end, as compared to directly using the original perturbed speech. Visualization of the features of the laughter perturbed speech and those generated by the proposed front-end further demonstrates the effectiveness of our approach.

Sri Harsha Dumpala, Rupayan Chakraborty, Sunil Kumar Kopparapu

Deep learning based discriminative methods, being the state-of-the-art machine learning techniques, are ill-suited for learning from lower amounts of data. In this paper, we propose a novel framework, called simultaneous two sample learning (s2sL), to effectively learn the class discriminative characteristics, even from very low amount of data. In s2sL, more than one sample (here, two samples) are simultaneously considered to both, train and test the classifier. We demonstrate our approach for speech/music discrimination and emotion classification through experiments. Further, we also show the effectiveness of s2sL approach for classification in low-resource scenario, and for imbalanced data.

Sri Harsha Dumpala, Ashish Panda, Sunil Kumar Kopparapu

Conversational speech not only contains several variants of neutral speech but is also prominently interlaced with several speaker generated non-speech sounds such as laughter and breath. A robust speaker recognition system should be capable of recognizing a speaker irrespective of these variations in his speech. An understanding of whether the speaker-specific information represented by these variations is similar or not helps build a good speaker recognition system. In this paper, speaker variations captured by neutral speech of a speaker is analyzed by considering speech-laugh (a variant of neutral speech) and laughter (non-speech) sounds of the speaker. We study an i-vector-based speaker recognition system trained only on neutral speech and evaluate its performance on speech-laugh and laughter. Further, we analyze the effect of including laughter sounds during training of an i-vector-basedspeaker recognition system. Our experimental results show that the inclusion of laughter sounds during training seem to provide complementary speaker-specific information which results in an overall improved performance of the speaker recognition system, especially on the utterances with speech-laugh segments.

Sri Harsha Dumpala, Rupayan Chakraborty, Sunil Kumar Kopparapu

Recurrent neural network (RNN) are being extensively used over feed-forward neural networks (FFNN) because of their inherent capability to capture temporal relationships that exist in the sequential data such as speech. This aspect of RNN is advantageous especially when there is no a priori knowledge about the temporal correlations within the data. However, RNNs require large amount of data to learn these temporal correlations, limiting their advantage in low resource scenarios. It is not immediately clear (a) how a priori temporal knowledge can be used in a FFNN architecture (b) how a FFNN performs when provided with this knowledge about temporal correlations (assuming available) during training. The objective of this paper is to explore k-FFNN, namely a FFNN architecture that can incorporate the a priori knowledge of the temporal relationships within the data sequence during training and compare k-FFNN performance with RNN in a low resource scenario. We evaluate the performance of k-FFNN and RNN by extensive experimentation on MediaEval 2016 audio data ("Emotional Impact of Movies" task). Experimental results show that the performance of k-FFNN is comparable to RNN, and in some scenarios k-FFNN performs better than RNN when temporal knowledge is injected into FFNN architecture. The main contributions of this paper are (a) fusing a priori knowledge into FFNN architecture to construct a k-FFNN and (b) analyzing the performance of k-FFNN with respect to RNN for different size of training data.