Saikat Chakraborty, University of Virginia, Charlottesville, VA, USA.
Title: Tree2Tree Neural Translation Model for LearningSource Code Changes
Abstract: The way developers edit day-to-day code tend to be repetitive and often use existing code elements. Many researchers tried to automate this tedious task of code changes by learning from specific change templates and applied to limited scope. The advancement of Neural Machine Translation (NMT) and the availability of the vast open source software evolutionary data open up a new possibility of automatically learning those templates from the wild. However, unlike natural languages, for which NMT techniques were originally designed, source code and the changes have certain properties. For instance, compared to natural language source code vocabulary can be virtually infinite. Further, any good change in code should not break its syntactic structure. Thus, deploying state-of-the-art NMT models without domain adaptation may poorly serve the purpose. To this end, in this work, we propose a novel Tree2Tree Neural Machine Translation system to model source code changes and learn code change patterns from the wild. We realize our model with a change suggestion engine: CODIT. We train the model with more than 30k real-world changes and evaluate it with 6k patches. Our evaluation shows the effectiveness of CODIT in learning and suggesting abstract change templates. CODIT also shows promise in suggesting concrete patches and generating bug fixes.
Bio: I received the B.Sc. degree in computer science and engineering from the Bangladesh University of Engineering and Technology, Dhaka, Bangladesh, in 2014. I am currently pursuing the Ph.D. degree with the University of Virginia, Charlottesville, VA, USA. I work with Professor Baishakhi Ray. I work on the intersection of Artificial Intelligence and Software Systems. My research interest is on designing intelligent tools and models to extract information from big sources of software information (e.g. Codebase), design and develop analysis tool to aid developer for better maintenance of software. I am interested in developing intelligent systems that help reduce human effort to detect, locate and fix bugs through intelligent program analysis and machine learning. I collaborate with research groups from UCLA, MSR UK, Fujitsu Lab etc.
Md Mahmudul Hasan, PhD candidate, Department of Mechanical and Industrial Engineering, Northeastern University
Title: Addressing Massachusetts’s Opioid Crisis: An integrated approach of Data Mining and Machine Learning
Abstract: In recent years, the addiction and abuse of prescription opioids has reached an epidemic level in the U.S, causing the drug overdose-related deaths to be escalated by five times in 2017 as compared to 1999—the largest proliferation in overdose related death ever recognized in this country’s history. In the state of Massachusetts, a similar scenario has been unfolded, even in a more alarming rate with the average death rate associated with prescription opioids surpassing the overall nation’s average. Given this public health emergency, immediate actions are required to make communities sufficiently resilient against the opioid addiction epidemic. To this end, we primarily focused on detecting the variation of opioids prescribing pattern of popularly prescribed opioid analgesics in different geographically dispersed regions of Massachusetts (economic regions, and rural versus urban areas) by leveraging the Massachusetts All Payer Claim Data (MA APCD) set. Results showed a recent increasing trends in the prescribing of suboxone—an opioid largely prescribed for managing opioid addiction, and clinical circumstances in which it is prescribed. Further, we are studying the pattern of long-term opioid usage to reveal the underlying factors that lead a naïve opioid patient to develop dependency on those drugs. Leveraging these influential risk factors, we intend to develop a risk prediction model to determine a patient’s likelihood of developing dependency on opioids. The proposed approach is expected to tackle the ravaging opioid crisis, partly in ensuring the judicious use of prescription opioids for patients who genuinely require them, and holistically deterring the flow of excess opioids to secondary users, leading to enhanced community resilience against this recent public health crisis.
Bio: Md Mahmudul Hasan is a PhD candidate in the department of Mechanical and Industrial Engineering at Northeastern University. Prior to that he has received his M.Sc. and B.Sc. in Industrial and Production Engineering department at Bangladesh University of Engineering and Technology (BUET). His research interest encompasses the application of advanced data-mining and machine learning techniques in solving the assorted problems of U.S. healthcare system by leveraging the big-data analytics. His recent work intends to develop a Decision Support System using Massachusetts All Payer Claim Datasets (MA APCD) to assist health care practitioners to prescribe opioid analgesics in a judicious way in an effort to reduce the likelihood of long term dependencies on and being overdosed by opioids. With the help such research integrated with other interdisciplinary approach his vision is to enhance community resilience against the crisis of nation-wide epidemic of opioid addiction and overdose.