20 accepted papers
While large language models excel in a variety of natural language processing (NLP) tasks, to perform well on spoken language understanding (SLU) tasks, they must either rely on off-the-shelf automatic speech recognition (ASR) systems for transcription, or be equipped with an in-built speech modalit…
Device-directed speech detection (DDSD) is the binary classification task of distinguishing between queries directed at a voice assistant versus side conversation or background speech. State-of-the-art DDSD systems use verbal cues, e.g acoustic, text and/or automatic speech recognition system (ASR)…
Streaming neural network models for fast frame-wise responses to various speech and sensory signals are widely adopted on resource-constrained platforms. Hence, increasing the learning capacity of such streaming models (i.e., by adding more parameters) to improve the predictive power may not be viab…
Accurate prediction of the user intent to interact with a voice assistant (VA) on a device (e.g. a smartphone) is critical for achieving naturalistic, engaging, and privacy-centric interactions with the VA. To this end, we present a novel approach to predict the user intention (whether the user is s…
Deep Neural Network-based speech recognition systems are widely used in most speech processing applications. To achieve better model robustness and accuracy, these networks are constructed with millions of parameters, making them storage and compute-intensive. In this paper, we propose Bin & Quant (…
In this paper we demonstrate speech synthesis using different electroencephalography (EEG) feature sets recently introduced in [1]. We make use of a recurrent neural network (RNN) regression model to predict acoustic features directly from EEG features. We demonstrate our results using EEG features…
The fast temporal changes of a millimeter wave channel necessitate frequent estimation of the channel. Power reduction techniques for the channel estimation process for ultra-wideband 5G systems are highly desirable. High speed analog-to-digital converters for the wideband data conversion and high s…
Understanding the neural processing of natural speech processing is an important first step for designing Brain-Computer Interface (BCI) based speech enhancement and speech recognition systems. Complex neural signals like electroencephalography (EEG) are time-varying and has a non-linear relationshi…
The performance of automatic speech recognition systems(ASR) degrades in the presence of noisy speech. This paper demonstrates that using electroencephalography (EEG) can help automatic speech recognition systems overcome performance loss in the presence of noise. The paper also shows that distillat…
We present a blind collision resolution algorithm in slow fading channels based on retransmission diversity. The algorithm neither assumes packet nor symbol synchronization of the different users and it does not demand estimates of the arrival times of the colliding signals. The proposed scheme work…
Artificial intelligence is often touted as the ultimate automation technology capable of outperforming humans. It is also feared by some because of its potential to eliminate certain jobs. In this paper, we describe scenarios in which man-machine symbiosis, or properly designed combinations of man a…
A generalized sequential probability ratio test (GSPRT) is a classical algorithm for binary sequential hypothesis testing. Though it is well-studied in the literature, there has been no optimal design of this test due to the difficulty of choosing its thresholds. In this paper we formulate the binar…
Imaging of physiological functions using magnetic resonance imaging is limited due its slow data acquisition speed. Previously various techniques based on data sharing in the spatiotemporal k-space or sparse recovery methods have been proposed to increase imaging speeds in dynamic MRI. This paper pr…
Direction of arrival (DoA) estimation of high-resolution beams is critical for cell search and maintaining communication in millimeter wave (mmWave) cellular systems. All-digital solutions for DoA estimation, though desirable for their flexibility and performance, are impractical because of their us…
We consider the problem of selecting and ordering a subset of N' out of N observations to be presented to a human being in the context of a binary hypothesis testing problem. We restrict our attention to i.i.d. Gaussian observations. We propose an extension of the approximate subset sum algorithm, a…
We present a novel QRS complex detection scheme from ECG with motion artifact. The algorithm relies on subspace learning and template matching. QRS complex detection during exercise is a challenging problem because multiple artifacts affect the ECG measurement. Motion artifact is considered to be th…
In this paper, we consider the problem of optimally ordering information to a human subject to maximize detection performance in a binary hypothesis testing problem. We begin by proposing a modification of the traditional Bayesian solution to hypothesis testing problems to incorporate the effect of…
Chronic effects of electrode implantation in the brain tissue alter the neural channel signal-to-noise ratio (SNR) over time. Variability of signal quality over time poses a difficult challenge in long-term decoding of neural signals for Brain Computer Interface (BCI). Specifically, all channels obs…
In this paper we propose a novel detection algorithm that delivers impressive savings in the sensing time of the capacitive touchscreen systems using sequential energy detection methods. We also show that these savings can translate into a significant boost to the operational battery life of today's…
High spatial resolution in functional magnetic resonance imaging improves its sensitivity to brain activation signals by reducing partial volume effects. However, the long acquisition times required for high spatial resolution limit the temporal resolution in fMRI studies. Consequently, the low temp…