Tuka Alhanai

I build technologies that help improve health and wellbeing.

Welcome

My name is Tuka, and I'm a PhD candidate at the Massachusetts Institute of Technology. I apply techniques in signal processing, machine learning, and statistics to build A.I. systems that can decode speech into text, and interpret spoken interactions.

Media

Writers for the general public have an invaluable abilitiy to distill abstract and technical knowledge into an accessible and informative format. You can explore some of what they've written that relates to my work and activities.

Media

Can Artificial Intelligence Detect Depression in a Person’s Voice?
Nov 5th, 2018
Three Artificial Intelligence and Tech Tools Trying to Boost People's Mental Health
Oct 19th, 2018
MIT Technology Review Innovators Under 35 (MENA Region)
Sept 13th, 2018
MIT Researchers Teach a Neural Network to Detect Depression
Aug 29th, 2018
Model Can Detect More Naturally Detect Depression in Conversations
Aug 29th, 2018
MIT Researchers Find Doctors’ ‘Gut Feelings’ Play Key Role In Care Provision
July 25th, 2018
Doctors Rely On More Than Just Data for Medical Decision Making
July 20th, 2018
MassChallenge Boston Awards $1.5M in Equity-Free Prizes to Top Startups from its Eighth Cohort
Nov 3rd, 2017
Can an Algorithm Detect a Speaker's Mood?
Feb 23rd, 2017
Wearable AI Detects Tone Of Conversation To Make It Navigable (And Nicer) For All
Feb 6th, 2017
MIT Made A Wearable That Knows How a Conversation is Going
Feb 1st, 2017
Wearable AI System Can Detect a Conversation's Tone
Feb 1st, 2017

Signals, Patterns, and Understanding

I focus on problems that use a novel combination of information in order to develop a better understanding of how we speak, express emotions, and interact with each other. If we can understand this, then we are a step closer towards buidling a system we can use in our everyday lives to improve our health and wellbeing.

You can explore some examples of my work.

Projects

Detecting Depression from Interviews
A sequence-based neural network model that detects depression from audio and text transcripts of interviews.
Provider Sentiment and Resource Utilization
Decisions about patient care, especially in the ICU, are complicated. Doctors aren't machines, they also have "gut feelings" about their patients.
Role-specific Language Models
While audio is relatively easy to record, it remains a challenge to automatically diarize (who spoke when?), decode (what did they say?), and assess a subject’s cognitive health.
Spoken Language Biomarkers to Detect Cognitive Impairment
An automated system that evaluates speech and language features from audio recordings of neuropsychological exams.
Deep Learning for Speech-to-Text
Arabic speech-to-text system developed using 10 years of news data, and exploring a range of neural network topologies.
Lexical & Language Modeling for Arabic Speech-to-Text
Investigate lexical modeling, covering diacritics, pronunciation rules, and acoustic based pronunciation modeling.
Personalized Medication Dosing
A personalized medication dosing algorithm that is robust to missing data, and 29% more accurate than intensive care staff.
Images 2 Spreadsheet
An open-source tool for the transcription of paper-based medical spreadsheets.
Class - Speech & Language Technologies
A wearable A.I. that can interpret the mood of a spoken interaction.

Partnerships, Collaborations, and Experience

I've had the opportunity to work with some wonderful people and organisations. Their mentorship, guidance, and wisdom has allowed me to grow.

Mentors & Collaborators

Dr. Jim Glass
Dr. Rhoda Au
Ahmed Ali
Dr. Patrick Cardinal
Mohammad Ghassemi
Wei-ning Hsu
Mandy Korpusik
Sameer Khurana
Taha Merghani

Experience

Framingham Heart Study
In 1948, FHS scientists and participants embarked on an ambitious project to identify risk factors for heart disease. Today, the study remains a world-class epicenter for cutting-edge heart, brain, bone, and sleep research.
MIT Computer Science & Artificial Intelligence Laboratory
MIT's largest on-campus research insitute conducting work in theory, algorithms, robotics, language, vision, computational biology, and artificial intelligence.
Samsung Strategy & Innovation Center
SSIC’s focuses on strategic initiatives in areas such as digital health; data center and cloud computing; human-computer interface technologies; and the “Internet of Things,” among others.

Technology in the Wild

I like to see my work being used in real world applications, so I try hard to develop systems that can be deployed and used. I also have a startup - The Ghamut Corporation - that develops technologies at the interface of artificial intelligence and human interaction. Our projects include a mood sensing wearable (EMOTE) and a platform (CONNECT) to facilitate connections between prospective friends and co-workers. We also provide consulting services and end-to-end system development.

Products

Connect
A platform that facilitates face-to-face interactions. Meet the most interesting person you've never met. We also pay for your lunch ;-)
Emote - Estimated Mood of Two Entities
A wearable A.I. that can interpret the mood of a conversation.

Services

Databases
Web Development
Artificial Intelligence
App Development
Cloud Computing
Hardware

Awards

MIT Graduate Student Life Grant
MIT de Florez Fund
MindHandHeart Innovation Fund
Legatum Fellowship
Sandbox Innovation Fund
MassChallenge Winner