Data & Artificial Intelligence (AI)

Overview

According to a recent United Nations report, due to the digital revolution the volume of data in
the world is increasing exponentially. By some estimates, 90 percent of the data in the world
has been created in the last two years, and it is projected to increase by 40 percent annually. As
the amount of data in our society has proliferated, it fueled the growth of Artificial Intelligence (AI) which enables people to rethink how we integrate information, analyze data, and use the resulting insights to improve decision making.

The power of AI to solve intractable global and social problems is immense. Much like the impact it has had on the business community, AI can rapidly accelerate decision making, supply chain management, and technology development to improve the effectiveness of humanitarian and community relief efforts. Yet, the use of AI to solve major social challenges in areas of health care, transportation, energy, and education still lags, largely due to the “wicked” and complex nature of these problems.

The Challenge

The Data & AI track challenges teams to demonstrate how AI can improve, accelerate, and/or
streamline the solutions to major societal challenges. Proposals should clearly demonstrate
thoughtfulness across three dimensions: real-world impact, scalability, and ethics and safety.
Solutions may be domestic or international in scope and may focus on a wide range of topics,
including but not limited to areas: healthcare, transportation, energy, education, and national
security

The Data & AI track challenges teams to demonstrate how AI can improve, accelerate, and/or
streamline the solutions to major societal challenges. Proposals should clearly demonstrate
thoughtfulness across three dimensions: real-world impact, scalability, and ethics and safety.
Solutions may be domestic or international in scope and may focus on a wide range of topics,
including but not limited to areas: healthcare, transportation, energy, education, and national
security

Past Winners and Examples

Note: This is the first year Big Ideas will offer an “AI” track. The winners listed below are projects which won in other tracks, but would have been eligible for AI had it previously existed.

Examples of proposals include (but are not limited to) the following:

  • A platform that uses mobile data, augmented by AI and machine learning, to identify human trafficking patterns/routes.
  • A system that extracts traffic data to identify common choke points and recommend new routes to ease congestion.
  • A wearable device that collects baseline health information and allows individuals to better predict health-related events and intervene.
  • A data analytics platform that enables disaster relief agencies to respond rapidly to humanitarian disasters such as earthquakes, and wildfires.
  • A mobile application designed to make the campus safer by analyzing crime patterns and recommending safe routes to students
  • A monitoring sensor and GPS locator that can detect where and when in the supply chain agricultural produce spoils.
2020 - EdVisor
fakenet-ai
2020 - FakeNetAI
BearCare
2017 - BearCare
2015 - USeeData

EdVisor

Team Members:

School:

UC Berkeley

EdVisor is pioneering equitable access to higher education, and we understand that many aspiring community college students cannot commit as much time to planning their academic journey. Higher education in the United States has become less accessible to underserved ethnic minorities, immigrants, and those from socio economically depressed communities. While there are many factors that contribute to the cumulative disadvantages in equity within our country, education is among the most evident and consequential. Today, EdVisor is leveraging the latest technology in data analytics, machine learning, and software development to simplify and optimize the community college degree planning process. EdVisor will allow students the unbiased freedom to plan faster and pivot seamlessly in their education with full transparency and support, resulting in increased enrollment, retention, and graduation rates.

FakeNetAI

Team Members:

School:

UC Berkeley

Over one billion hours of video are uploaded to the internet daily. Deepfakes, videos manipulated using deep learning techniques, represent a tiny fraction of those videos but are growing rapidly, doubling in the past nine months. Deepfake quality is improving to the point where the best are unidentifiable by human reviewers–and already have been used for nefarious purposes, including inserting a person into pornography or manipulating politics. As a result of the growing quality and ability to insert Deepfakes into a sea of uploaded content, the risk of their spreading and causing damage for content hosting companies is increasing. It is critical to be able to identify and react to Deepfakes by flagging or removing them. FakeNetAI’s Deepfake detection technology enables content hosting platforms to detect Deepfake content to counteract this growing threat. FakeNetAI’s technologies allow companies to respond to the threat even as Deepfakes continue to grow exponentially.

NeoMotion AI

Team Members:

School:

UC Berkeley, UC San Francisco

Stroke rehabilitation is often inaccessible, expensive, and requires a lot of scarce, highly trained professionals. By harnessing the processing power of smartphones in combination with recent advances in artificial intelligence, NeoMotion AI will be able to improve rehabilitation at a worldwide scale. NeoMotion AI is run on AI-based pose estimation algorithms and it optimizes them for usage on smartphones without internet connection. Using a smartphone camera and this software, the solution involves tracking the coordinates of every joint of the upper and lower limbs, proving patients, rehabilitation specialists, and physicians with a tool to track patient’s rehabilitation progress over time. At a later stage, performing rehabilitation exercises in front of a smartphone would allow patients to receive personalized exercise corrections or new and adaptive exercises suggestions. NeoMotion AI can provide an engaging rehabilitation experience for stroke patients through a social platform, creating a sense of community, and a more integrated management system for physical therapists or physicians.

BearCare

Team Members:

School:

UC Berkeley

With 38,204 matriculated students and thousands of programs available on campus, the University of California Berkeley can often be a daunting place to find health services and support. In response to the issues that students face, the American Medical Student Association: Community and Public Health Committee seeks to create an iOS mobile phone application that works as a unique one-stop resource for addressing health at Berkeley connecting students with available support. While there are some other applications which provide limited information or help, the unique point of this application is to consolidate all resources that students would need into one centralized location. The application addresses mental health, sleep, nutrition, exercise, sexual health, illnesses, and provides assistance when walking home or at a party situation. In addition to direct student benefits, anonymous data from all application users could eventually be collected and analyzed for use with future campus projects.

USeeData

Team Members:

School:

UC Berkeley

As a field, environmental science has not yet had the impact it needs to have on the general public, mostly due to limited amount of exposure to information. A powerful visualization tool would help bring all types environmental causes to the forefront and help the general population understand how the environment influences their day to day lives. This project seeks to achieve that goal by creating an open source environmental data visualization suite for researchers and data scientists. The tool set will be especially tailored toward geographical and energy related data; however, the end project will create a much more versatile set of abilities. This tool would make it incredibly simple for researchers to submit any kind of geographic data and create meaningful visualizations of that data, without the need for much technical knowledge on the part of the researchers.

SEPT
2020

18

Come get inspired by Big Ideas alumni and their stories of how they transformed their early-stage idea into a successful social venture.