X-Lab Projects

X-Lab Projects provide real-world experiential learning and/or research translation. X-Lab projects also build upon the Innovation Collider model as well as the Innovation Engineering framework.


Contact Ikhlaq Sidhu for all project information at sidhu@berkeley.edu, please add a statement of interest and CV if you are interested in opportunities related to these projects.

Discovery Project: AI / Social Reliability and Ethics Estimator:

This AI and algorithm driven project will develop an approach and system to estimate if a person or organization can be trusted. How can we know if people at Facebook or Disney can be trusted? The same question could be asked for a job candidate. Today, we call random references which actually might be subjective. In China, a social network based estimator has been created to augment credit score. Similarly, it is possible that the trust level of a person's social network has influence on their own trust and ethics level. In this project, the team will research the different approaches and develop a distributed algorithm based approach to estimate social trust, reliability, and even ethics.

Discovery Project: AI Super-Recruiter

This project is about using a data science algorithm to emulate a professional recruiter in their job of finding good candidates for executive positions. A small team of students will work with a job recruiting firm in Asia. The firm will have one or more of their top recruiters mark resumes to explain what they look for in a set of 20 or more CVs for executive positions. By understanding what they look for, the team will develop an algorithm and/or rule based approach to selecting resumes. Results will be shown to the executive job search firm. If the project continues, there will be option for funded or paid work in the next phase of the project. The initial project phase should take 2 months to complete as a team-oriented independent study.Qualified students should data science and pr machine learning background and are interested to apply this capability in this real life research project.
Data Center Energy Efficiency [DRAFT]

Discovery Project: Server Farm Energy Saver

This project will last one month and will kick off as soon as a team of 3 people have been formed. The goal of the project is to do background research to understand how a Data Center can best save energy. The output will be slide presentation to be shared with selected data centers as part of a proposal for a funded project. If the project achieves funding, then there will be additional paid opportunities for that student team and potentially others.
Key issue: Data centers with moderate sophistication use very large amounts of energy which makes it difficult to compete. Data centers are interested to understand the best methods to reduce energy usage. To do this, a student team can investigate 1) best practices of current data centers, 2) what controls or options are available to most data centers to reduce energy depending on demand, 3) can data/AI algorithms be used to solve this problems, 4) how much energy savings can be expected using different types of strategies.
See the initial project results at the slides to the left.<<<

Honda Smart Park Project

Students will work on a business model and technology for a smart parking application using the innovation engineering framework. In a past semester, a student team in Data-X, developed a first version of the application. The team used data from Honda to estimate whether cars in the area were driving or parking. Combining this information with Google map APIs, the application could tell a driver where to park, and how to walk to the final destination. In the next version of the parking application, we want to know other sources of data for parking lot information. How would the application work with Honda and also earn revenue. (This topic is related to the SCET's Amazoogle course). And of course, the team will implement the application.

Volvo SLAM Project

The Volvo Slam project is a technically sophisticated, research project related to autonomous driving, mapping, and mobility algorithms. It is intended to find a way to track the position of a vehicle at slow speeds when GPS signal is NOT available. This information may be compensated by accelerometer readings and/or video and photos taken with timestamps. The core of the problem is drift of the estimation of the location. A common application would be to know the car's location even when it is a parking garage without satellite signals. Technical skills required: One more of CS application development, data/ai, control theory, video/image/signal processing