About me
I am currently a PhD candidate at Purdue University, specializing in Digital Forensics and Data Science. My research focuses on leveraging AI to uncover criminal behavior patterns, with a particular emphasis on using machine learning (ML) and Natural Language Processing (NLP) to generate actionable crime insights for law enforcement investigations.
After more than nine years working as a forensic scientist in Brazil, I observed that many criminals carefully plan their offenses, revealing patterns in their behavior known as Modus Operandi (MO). Despite advancements in AI, identifying these patterns often still relies heavily on manual investigative work. I am deeply interested in how recent technological breakthroughs, particularly large language models (LLMs), can help analyze complex criminal data and enhance investigative efforts. My research spans both physical and cyber crimes, and I am also interested in exploring the ethical implications of applying ML and LLMs within the criminal justice system.
Just as significant advancements in healthcare have emerged through strong collaborations between academics and practitioners, I believe the criminal justice field will experience similar progress over the next 5-10 years. Crimes (especially cyber crime) are becoming increasingly complex, requiring interdisciplinary and innovative solutions to effectively address and combat them.
Collaboration from all backgrounds are welcome. Please reach out if you are intereseted in these or related topics.
Selected Papers:
-
Health Disparities through Generative AI Models: A Comparison Study Using A Domain Specific large language model
Yohn Jairo Parra Bautista, Vinicious Lima, Carlos Theran, Richard Alo. Proceedings of the Future Technologies Conference, 2023.
-
Identifying Risk Patterns in Brazilian Police Reports Preceding Femicides: A Long Short Term Memory (LSTM) Based Analysis
Vinicius Lima, Jacque Almeida de Oliveira. 2023 IEEE Global Humanitarian Technology Conference (GHTC).
-
Forensic intelligence approach on traffic accidents in the brazilian federal district
Vinícius de Oliveira Lima, José Marcos VA Gois. WIT Transactions on the Built Environment.
-
Hotspot Prediction of Severe Traffic Accidents in the Federal District of Brazil
Vinicius Lima, Vetria Byrd. SMART 2023, The Twelfth International Conference on Smart Cities, Systems, Devices and Technologies.
Other Projects:
-
Patterns in Violence Against Women
I developed an LSTM model to predict sequential behavior in severe cases of violence against women.
-
Machine Learning from Scratch
In this project, I built a Logistic Regression and a Neural Network from scratch with the minimal use of libraries.
-
Traffic Accident Hotspots Prediction
I tested a few ML tecniques to predict the number of accidents in regions of the Federal District of Brazil.
-
Sentiment Analysis of Movie Reviews
I tested a basic NLP tecnique to automatically classify IMdB reviews (positive or negative comments).
Other Accomplishments:
