Finding the Key to Success: Computer Science PhD Candidate Bhaskar Ghosh Helps AI Algorithms Improve Their Performance
As a computer science PhD candidate studying artificial intelligence (AI) and machine learning, Bhaskar Ghosh is working to make technology run more efficiently.
小蝌蚪APP淎rtificial intelligence is everywhere right now, starting from your phone to your car. But, the work of each AI function is different. We generally need to choose a good algorithm to use for your AI task. My research can be used to help predict the best algorithm to use for your AI application without actually having to run multiple iterations,小蝌蚪APP he says.
Bhaskar is preparing to graduate in Fall 2021. He is completing his dissertation under the direction of Dr. Michael Totaro, associate professor in the School of Computing and Informatics.
小蝌蚪APP淭he main question that I have asked throughout my work was, how do I make this process faster without having to make any changes to the existing architecture? I小蝌蚪APP檝e applied this idea to other AI algorithms to help them perform faster,小蝌蚪APP he explains.
小蝌蚪APP淎s an analogy, think if you have a bunch of keys and one lock. My work would give you a good estimate of which key to use to open the lock.小蝌蚪APP
Continuing the Journey
Bhaskhar completed a master小蝌蚪APP檚 degree in Computer Science from UL Lafayette and received an offer from the School of Computing and Informatics to continue his research through the doctoral program.
小蝌蚪APP淭his was an opportunity I could not decline,小蝌蚪APP he recalls. 小蝌蚪APP淕etting to work at the 小蝌蚪APP that I love and stay in the city that has grown dear to me was something that I did not want to give up.小蝌蚪APP
As technology features evermore heavily in our everyday lives, his studies have shown him the importance of understanding how systems operate.
小蝌蚪APP淭here are actually quite a few things that I have learned from my research that are very important but hardly spoken about,小蝌蚪APP he says.小蝌蚪APP淭his was very shocking for me to learn about the easy access to data that people willingly provide the services or devices that they use. The importance of data and also the insecurity of this data has been an eye opener for me.小蝌蚪APP
During his time at UL Lafayette, Bhaskar has learned to effectively communicate with the public about his research findings. His Three Minute Thesis (3MT庐) presentation 小蝌蚪APP淎re You Certain It's Time to Prepare?小蝌蚪APP used the example of hurricane preparation to discuss the importance of effectively using algorithms and historic data to create weather predictions.
Creating forecasts that are accurate and trustworthy, he argues, are essential to motivating people to prepare appropriately for natural disasters. Bhaskar小蝌蚪APP檚 presentation was chosen as the People小蝌蚪APP檚 Choice winner in the UL Lafayette Graduate School 2020-2021 3MT庐 Competition.
Building Resilience
The path to completing a PhD requires an enormous amount of time and effort, but also the ability to take failure in stride, Bhaskar explains.
小蝌蚪APP淵our tasks might not always be successful, but you must look ahead, learn from it, and move on. I would say getting a PhD consists of 70% rejections and failed experiments, but that 30% of success that you achieve is something that must be cherished. You have reached this because you have failed 70% of the time and the learnings from those failures is what sets you apart in that field.小蝌蚪APP
This hard-earned lesson to maintain optimism and keep moving forward is something Bhaskar will take with him as he graduates and continues his research at a university or in an industry position.
小蝌蚪APP淭he balancing act of managing the stress of failures and never giving up has been the most valuable lesson that I have learned,小蝌蚪APP he says.
Having the support of a proactive scholarly community on campus has also been key to his success.
小蝌蚪APP淚 would certainly motivate any student who is considering joining UL Lafayette for their PhD,小蝌蚪APP he says. 小蝌蚪APP淭he faculty, staff and students are very helpful. They are always there to go above and beyond.小蝌蚪APP
