arXiv subjects

Prateek Verma

Publications and source records attributed to Prateek Verma.

43 records · Page 3Linked to original sources

arXiv2019

End-to-End Spoken Language Translation

In this paper, we address the task of spoken language understanding. We present a method for translating spoken sentences from one language into spoken sentences in another language. Given spectrogram-spectrogram pairs, our model can be trained completely from scratch to translate unseen sentences. Our method consists of a pyramidal-bidirectional recurrent network combined with a convolutional network to output sentence-level spectrograms in the target language. Empirically, our model achieves competitive performance with state-of-the-art methods on multiple languages and can generalize to unseen speakers.

cs.CL↗

arXiv2019

Neuralogram: A Deep Neural Network Based Representation for Audio Signals

We propose the Neuralogram -- a deep neural network based representation for understanding audio signals which, as the name suggests, transforms an audio signal to a dense, compact representation based upon embeddings learned via a neural architecture. Through a series of probing signals, we show how our representation can encapsulate pitch, timbre and rhythm-based information, and other attributes. This representation suggests a method for revealing meaningful relationships in arbitrarily long audio signals that are not readily represented by existing algorithms. This has the potential for numerous applications in audio understanding, music recommendation, meta-data extraction to name a few.

cs.SD↗

arXiv2019

Audio-Linguistic Embeddings for Spoken Sentences

We propose spoken sentence embeddings which capture both acoustic and linguistic content. While existing works operate at the character, phoneme, or word level, our method learns long-term dependencies by modeling speech at the sentence level. Formulated as an audio-linguistic multitask learning problem, our encoder-decoder model simultaneously reconstructs acoustic and natural language features from audio. Our results show that spoken sentence embeddings outperform phoneme and word-level baselines on speech recognition and emotion recognition tasks. Ablation studies show that our embeddings can better model high-level acoustic concepts while retaining linguistic content. Overall, our work illustrates the viability of generic, multi-modal sentence embeddings for spoken language understanding.

cs.SD↗

arXiv2018

Towards a simplified ontology for better e-commerce search

Query Understanding is a semantic search method that can classify tokens in a customer's search query to entities such as Product, Brand, etc. This method can overcome the limitations of bag-of-words methods but requires an ontology. We show that current ontologies are not optimized for search and propose a simplified ontology framework designed specifically for e-commerce search and retrieval. We also present three methods for automatically extracting product classes for the proposed ontology and compare their performance relative to each other.

cs.IR↗

arXiv2018

Conditional End-to-End Audio Transforms

We present an end-to-end method for transforming audio from one style to another. For the case of speech, by conditioning on speaker identities, we can train a single model to transform words spoken by multiple people into multiple target voices. For the case of music, we can specify musical instruments and achieve the same result. Architecturally, our method is a fully-differentiable sequence-to-sequence model based on convolutional and hierarchical recurrent neural networks. It is designed to capture long-term acoustic dependencies, requires minimal post-processing, and produces realistic audio transforms. Ablation studies confirm that our model can separate speaker and instrument properties from acoustic content at different receptive fields. Empirically, our method achieves competitive performance on community-standard datasets.

cs.SD↗

arXiv2018

Bribery Games on Interdependent Complex Networks

Bribe demands present a social conflict scenario where decisions have wide-ranging economic and ethical consequences. Nevertheless, such incidents occur daily in many countries across the globe. Harassment bribery constitute a significant sub-set of such bribery incidents where a government official demands a bribe for providing a service to a citizen legally entitled to it. We employ an evolutionary game-theoretic framework to analyse the evolution of corrupt and honest strategies in structured populations characterized by an interdependent complex network. The effects of changing network topology, average number of links and asymmetry in size of the citizen and officer population on the proliferation of incidents of bribery are explored. A complex network topology is found to be beneficial for the dominance of corrupt strategies over a larger region of phase space when compared with the outcome for a regular network, for equal citizen and officer population sizes. However, the extent of the advantage depends critically on the network degree and topology. A different trend is observed when there is a difference between the citizen and officer population sizes. Under those circumstances, increasing randomness of the underlying citizen network can be beneficial to the fixation of honest officers up to a certain value of the network degree. Our analysis reveals how the interplay between network topology, connectivity and strategy update rules can affect population level outcomes in such asymmetric games.

physics.soc-ph↗

arXiv2018

Neural Style Transfer for Audio Spectograms

There has been fascinating work on creating artistic transformations of images by Gatys. This was revolutionary in how we can in some sense alter the 'style' of an image while generally preserving its 'content'. In our work, we present a method for creating new sounds using a similar approach, treating it as a style-transfer problem, starting from a random-noise input signal and iteratively using back-propagation to optimize the sound to conform to filter-outputs from a pre-trained neural architecture of interest. For demonstration, we investigate two different tasks, resulting in bandwidth expansion/compression, and timbral transfer from singing voice to musical instruments. A feature of our method is that a single architecture can generate these different audio-style-transfer types using the same set of parameters which otherwise require different complex hand-tuned diverse signal processing pipelines.

cs.SD↗