My research interest lies in the area of speech processing and machine learning oriented in biomedical applications and human health. I am particularly interested in tackling problems of speech and language processing that is motivated by practical needs in the biomedical and health domain. My current projects focus specifically on cutting-edge methods to screen for speech and/or cognitive disorders using data collected and interpreted by a computer. I have developed several new programs and tools for monitoring Parkinson’s disease, Alzheimer’s disease, Autism Spectrum Disorder (ASD), and Mild Cognitive Impairment (MCI); created devices that are better suited to distinguish between the human voice and background noise; and discovered more effective methods to accurately and robustly extract speech features employed in these applications. My conducted research and expertise addresses the need for early detection and monitoring of the above-mentioned disorders and uses the manifestations that are observed in patients’ speech to offer a wealth of insights.
Courses: Deep Learning, Automatic Speech Recognition
I study ways to apply natural language processing techniques to a variety of biomedical problems, including clinical information retrieval, secondary analysis of electronic medical record data, and automatic assessment of neuropsychological disorders. My current projects include an automated tool for analyzing medical articles for use in systematic reviews, a novel communication platform for individuals with primary progressive aphasia, and an ongoing study of the linguistic patterns of children with autism.
Courses: Natural Language Processing, Information Retrieval, Problem Solving with Large Clusters, Ethics for CS/EE
I am interested in the foundations of social language development, Autism Spectrum Disorder (ASD) assessment and diagnosis, and the intersection of language and social communication. As a speech-language pathologist, I participate in a multidisciplinary assessment clinic for ASD at the Child Development & Rehabilitation Center. My current research projects include utilizing a general social-emotional screener as an ASD specific tool, investigating the developmental trajectories of children born prematurely, and developing an automated, voice-based assessment of children’s language abilities. My work draws upon a background in linguistics, psychology, speech and hearing science, special education, and pediatric healthcare.
Courses: Ethics for CS/EE
My research area is dialogue management, and models of the conversational aspects of speech, such as turn-taking, discourse markers, disfluencies, and stuttering. I conduct empirical studies to better understand how people engage in these activities, and build computation models, for understanding what people are doing, and for building better spoken dialogue systems. My current projects include understanding how people engage in turn-taking, so that there is little overlap or delay in conversation; understanding how people with mild cognitive disorder manage misunderstandings, and building tools, based on automatic speech recognition, that can assist clinicians in treating people who stutter.
Courses: Automata and Formal Languages, Dialogue, Artificial Intelligence
I study the analysis, processing, and synthesis of biological signals, with a focus on human speech. The results of my speech research lead to improved scientific understanding about speech production and perception with regards to intelligibility, and to technical innovations that may improve the quality of life for individuals with speech disorders in the future. My current projects include creating a computer-based pronunciation analysis system for children with speech sound disorders, enhancing a Text-to-Speech system to allow it to mimic its operator's voice, and developing an unobtrusive screening method for sleep apnea based on sleep sounds.
Courses: Data Science Programming, Speech Synthesis, Speech Signal Processing
As a developmental psychologist, I study the development of core symptoms and comorbid problems in children with autism spectrum disorders (ASD). I conduct empirical studies and secondary data analyses to better understand the mechanisms underlying both social and health-related difficulties in this population. My current research focuses on defining and quantifying the language phenotype in ASD, analyzing the prevalence and factors associated with overweight and obesity in children with ASD, and understanding the epidemiology of ASDs.
Courses: Probability & Statistics, Empirical Research Methods
My research areas are in machine learning, image processing and analysis, and computer vision. I develop algorithms that can analyze large amounts of data, usually in many dimensions and from heterogeneous sources, for the purposes of quantification, classification, prediction and motion tracking. My current projects include motion tracking to study cardiac mechanics using 4D echocardiography, characterizaiton of MRI images for accurate diagnosis and therapy response assessment of breast cancer, and characterization of motion using accelerometers for various movement disorders.
Courses: Machine Learning, Introduction to Image Processing
I draw tools from statistical signal processing, systems theory, information theory, machine learning, pattern recognition, and neural computation in working on problems in:
Computational Biology: I am interested in discovery and implementation of algorithms that facilitate the understanding of biological processes. Particular applications have included pathway modeling, sequence-to-function analysis of genes and proteins, ontology/schema development for biological databases (http://www.biospice.org), and determination of sleep/arousal states in rodents from electrophysiological recordings.
Statistical Signal Processing and Automatic Speech/Speaker Recognition: My work has concentrated on automatic speech and speaker recognition, especially robust speech recognition, modeling of prosody, and multiresolution speech representations. Earlier work has included information theoretic bases for estimation, and stochastic processes.
Jan van Santen
I create new algorithms for the analysis of biological signals, specifically human speech (in particular, prosody), language, and gesture (using accelerometers), as well as animal vocal emissions. This work is increasingly focused on computerized behavioral assessment and intervention for neurological disorders, but also involves general-purpose systems such as new prosody generation engines for text to speech (TTS) synthesis. My current projects include language in autism spectrum disorders, prosody generation for TTS, prosody in Parkinson's disease, atypical movement in infants at risk for autism spectrum disorders, and (via our spin-off, BioSpeech, Inc.) auditory rehabilitation, analysis of mouse model vocal emissions, and phonemic awareness training.
I'm interested in meaning. How can we understand text or speech at the word-level? Sentence-level. Document-level? I've found that oftentimes, meaning must be understood with respect to domain context, so I work on the medical context of millions of clinician's notes — information retrieval, information extraction, and corpus analysis. I've also found that the meaning of language reflects its situational context, and I must account for this if I want my work to have any real-world impact. This drives my research in modeling the medical history of patients from a holistic patient-level perspective; it also informs the NLM-funded project that I lead, on the practical problem of retrieving patient cohorts from clinical text.
Courses: Analyzing Sequences, Natural Language Processing