Projects

Adaptive Prosthetics Program

Principal Investigators
Patrick M. Pilarski, Richard S. Sutton

Intelligent Artificial Limbs & Biomedical Devices

The Adaptive Prosthetics Program, a collaboration between Amii and the BLINC Lab, is an interdisciplinary initiative focused on real-time machine learning methods for assistive rehabilitation and intelligent artificial limbs. Through the development of fundamental algorithms and the translation of methodology into clinical benefit, the program seeks to increase patients’ ability to customize and control assistive biomedical devices.

The Adaptive Prosthetics Program explores fundamental and applied methods for real-time prediction, adaptive control and direct human-machine interaction. Technologies developed through the program include the Bento Arm and the HANDi Hand, both of which have open-sourced hardware and software through the BLINCdev community.

See Also

Development of the Bento Arm: an Improved Robotic Arm for Myoelectric Training and Research” published at MEC ’14: Myoelectric Controls Symposium

Development of the HANDi Hand: an Inexpensive, Multi-Articulating, Sensorized Hand for Machine Learning Research in Myoelectric Control” published at MEC ’17: Myoelectric Controls Symposium

Projects

Arcade Learning Environment

Principal Investigator
Michael Bowling

Atari 2600 Platform for General Artificial Intelligence Development

The Arcade Learning Environment (ALE) is a software framework designed to facilitate the development and testing of general AI agents. ALE was created as both a challenge problem for AI researchers and a method for developing and evaluating innovations and technologies.

Historically, many AI advancements have been developed and tested through games (e.g. CheckersChessPoker and, most recently, Go), which offer controlled, well-understood environments with easily defined measures for success. Games also give researchers a concrete and relatable way to demonstrate artificial intelligence to a broad audience.

The Arcade Learning Environment, powered by the Stella Atari Emulator, provides an interface to hundreds of Atari 2600 games. These diverse game environments are complex enough to be challenging for an AI agent to solve yet simple enough to enable progress.

ALE is available to researchers and hobbyists alike with Atari now being used by groups like Google DeepMind to develop and test their deep reinforcement learning methodologies.

See Also

The Arcade Learning Environment: An Evaluation Platform for General Agents” published in the Journal of Artificial Intelligence Research

Projects

Meerkat

Principal Investigators
Randy Goebel, Osmar Zaïane

Social Network Analysis & Visualization

Meerkat is an automated Social Network Analysis (SNA) tool used to analyze, visualize and interpret large or complex networks of information, allowing users to examine patterns and investigate relational dynamics.

The application uses information about the interactions between a set of objects (or nodes) within a network and lets the user employ different algorithms to automatically identify meaningful connections or highlight the most influential or central nodes in different ways.

Network analysis features include:

  • Automated community detection and analysis
  • Interactive visualization using general, community and metric-based layouts
  • Filtration and extraction of useful data
  • Dynamic analysis of network changes over time

Meerkat also provides tools for text mining, including polarity and emotion analysis, which give users the ability to examine text for positive and negative sentiments and a range of basic emotions. Meerkat ED, a version of the program that has been tailored specifically for educational environments, allows instructors to evaluate student activities in online discussion forums.

Projects

DeepStack

Principal Investigator
Michael Bowling

Problem we’re trying to solve

For several years, AI researchers have had a number of different techniques for predicting and planning optimal actions in situations of perfect information (where all actors have the same, full knowledge of the world). Techniques have been lacking for dealing with imperfect information situations (where actors do not have access to certain information or have access to information the other doesn’t). DeepStack seeks to successfully apply, for the first time, theoretical techniques for perfect information games into situations with imperfect information.

How will this help someone / an industry?

For computing scientists and AI researchers, DeepStack represents a foundational step forward in dealing with issues around predicting optimal actions in the face of ambiguity and uncertainty. The theoretical advancements demonstrated in DeepStack will open new avenues of research for scientists interested in building, and planning with, models of unknown, complex dynamic systems.

Type of MI used

Reinforcement learning, Deep learning

Projects

Diagnosing Tuberculosis

Principal Investigator:
Yutaka Yasui

Problem we’re trying to solve

Inexpensive, timely and accurate diagnosis of potential cases of Tuberculosis is of critical importance in regions of the world where resources are limited. Standard methods of diagnosis are often too expensive or resource intensive to be deployed in the very regions where tuberculosis is a significant problem. This leads to poor patient outcomes due to delayed treatment and undiagnosed illness.

How will this help someone / an industry?

The machine learning component of this work developed a new automated diagnostic method using image analysis which enables diagnosis that is more efficient and lower cost than standard methods, and because it is automated, poses lower biohazard risk to the technicians processing samples.

Partners

TV/HIV Research Foundation (Thailand)

Projects

FMRI-based Diagnosis & Treatment

Principal Investigator:
Russ Greiner

Problem we’re trying to solve

Current methods of diagnosing neurological and psychological disorders often rely on the subjective assessment of a patient’s symptoms by a clinical psychiatrist. These assessments can differ between psychiatrists, leading to different recommendations for a treatment plan. We aim to provide these psychologists with tools that provide objective criteria for diagnosis and the assessment of symptom severity in order to provide psychiatrists data-driven methodologies for assessing patients.

How will this help someone / an industry?

Our goal with this project is to use machine learning techniques to produce clinical tools that could assist medical doctors in providing faster, more effective treatment for neurological and psychiatric illnesses. We are exploring the use of brain imaging to diagnose mental disorders earlier and more accurately, to predict symptoms and their severity, and to predict which combination of drug therapies will work best for a given patient.

Partners

IBM Research

Projects

Patient-Specific Survival Prediction

Principal Investigator:
Russ Greiner

Problem we’re trying to solve

Prognostic modeling is an integral component in the treatment and management of patients. Currently being developed for the field of oncology, PSSP predicts individual survival distributions for patients from their electronic health record, significantly reducing the prediction error compared to the standard approach of using only the cancer site and stage.

How will this help someone / an industry?

More accurate survival time prediction can improve medical decision making (for example, by deciding whether a treatment option is cost-effective based on it’s added survival time, or by helping determine when a patient should be referred for end-of-life care.) The tool can be used more generally for any task that involved predicting a life-cycle (customer churn, diagnosing machine faults, etc).

Projects

Legal Reasoning

Principal Investigator:
Randy Goebel

Problem we’re trying to solve

A great deal of human resources are used in order to prepare for (or render a verdict in) a legal proceeding. The tools we are developing for information extraction and visualization in the legal domain extract legal concepts from text, identify chains of legal reasoning and answer questions by using textual entailment. Developing automated tools for information extraction and knowledge discovery will reduce the amount of effort and time needed for legal matters.

How will this help someone / an industry?

This project seeks to develop techniques for legal case reasoning, legal summarization and legal question answering in order to allow law practitioners to redirect their attention to tasks that require creativity or more complex reasoning.

Partners

National Institute for Informatics (Japan)

Type of MI used

Natural Language Processing, Information Extraction.

Projects

Ana – Automated Nanny Agent

Principal Investigator:
Osmar Zaïane

Problem we’re trying to solve

Ana, a conversational software agent (ie chatbot), is designed to converse with the elderly living at home to answer general questions and remind them of specific events. Ana is able to extract from conversations named entities (ie places, people, prescriptions, recipe names, etc.) as well as relationships (ie family tied, professions, activities, temporal events, etc.) Ana extracts information from text obtained from a speech-to-text converter; from this, it builds a personalized knowledge base that allows it to answer personal questions. Ana can also answer impersonal questions from sources on the Internet.

How will this help someone / an industry?

In addition to developing a speech interface for human-machine interaction, Ana seeks to improve elderly home care by providing a personal assistant and a digital companion. Ana helps with social needs (through questions and answers) and assists with simple home healthcare needs (ie. prescription reminders).

Type of MI used

Information Extraction, Natural Language Processing.

Projects

Advanced Analytics for Curling

Principal Investigator:
Michael Bowling

Problem we’re trying to solve

The Computer Curling Research Group focuses on deep analytics for the sport of curling for player analysis and system analysis, and to create tools that can translate AI discovered insights into improvements to human decision making.

How will this help someone / an industry?

The ultimate goal of this project is to develop tools and models that enable player/team assessment, strategic game modeling, and analytics for broadcast television.

Type of MI used

Deep learning, Search and Planning.

Projects

Intelligent Diabetes Management

Principal Investigator:
Russ Greiner

Problem we’re trying to solve

The current method of determining insulin dosages requires a patient to manually track their insulin levels multiple times a day, collect data over a certain period of time, present that data to a diabetologist, and have their dosage adjusted after weeks of using the wrong dose. This is delaying our ability to optimize treatment, depends on the patient’s commitment to tracking data, and requires a diabetologist to personally evaluate each case.

How will this help someone / an industry?

Machine learning is able to use patient data to adjust insulin levels in real time, making their treatment personalized, more accurate, and more affordable. It also increases the capacity for diabetologists to see more patients and help more people.

Partners

Alberta Diabetes Institute; top rated diabetologist from Alberta

Type of MI used

Reinforcement learning