News & Events

ICDS News

Above from left, Dana Calacci, assistant professor in the College of Information Sciences and Technology, and Enrico Casella, assistant professor of data science for animal systems in the College of Agricultural Sciences, have been named as ICDS co-hires.

ICDS welcomes Dana Calacci, Enrico Casella as co-hires

Posted on September 24, 2024

Editor’s Note: A version of this story was published on Penn State News.

 

UNIVERSITY PARK, Pa. — The Penn State Institute of Computational and Data Sciences (ICDS) recently welcomed two new co-hires to their staff. Dana Calacci, assistant professor in the College of Information Sciences and Technology, and Enrico Casella, assistant professor of data science for animal systems in the College of Agricultural Sciences, started their Penn State journeys this summer. 

“Dana and Enrico are the latest additions of co-hires in the domains of artificial intelligence and machine learning (AI/ML),” Dr. Guido Cervone, ICDS interim director, said. “They bring important new skills and complement an already robust cohort of co-hires that focus on AI/ML. That is a very important domain of research for the institute, one in which we have significantly invested in over the last few years, and one where we will continue investing in.” 

Calacci and Casella plan to use ICDS compute and engineering resources for their research that focuses on artificial intelligence (AI), computer vision and broader impacts on the communities in which their work serves. 

“I’m really happy to have ICDS resources, especially the GPUs,” Casella said. Besides, multidisciplinary work has always been my thing, and I have already experienced how easy it has been to connect with and network with the ICDS community.” 

 

Dana Calacci 

 

Calacci’s interdisciplinary work is inspired by her mission to make a real-world, positive impact on data ownership, its use, the incentive structures in which we collect and use data as well as the economies and markets that manage data’s value. 

While going to Massachusetts Institute of Technology pursuing a doctoral degree, Calacci was working on a variety of projects measuring human behavior quantitatively including deep reinforcement learning algorithms which were inspired by human networks, to trying to understand segregation in cities with mobile phone data.  

Calacci found a lack of regulation with large sets of sensitive data and has continued to think critically about how communities can gain more control over their data and use it to advance their interests, as well as better understand how artificial intelligence (AI) and data can harm communities as well. 

“We want to give communities the tools to understand how these technologies can impact them from day to day,” Calacci said. 

While at Penn State, Calacci is continuing the research she started within her doctoral program, in which she developed the Workers Algorithm Observatory (WAO), which has built tools for workers to crowdsource data about their work. 

Rideshare and delivery drivers and workers alike are “subject to opaque algorithmic systems that determine their pay and their working conditions,” Calacci said.  

“Data ownership and understanding how algorithms or data systems impact communities matters to workers,” Calacci said. “The data that gets collected about them has the potential to directly impact their working conditions, pay and things that impact their realities. Working with them [delivery drivers and rideshare drivers] has been a valuable way to understand, on the ground, how data systems impact people’s real experiences.” 

Calacci anticipates using ICDS resources to better understand how to build participatory tools, auditing and evaluating generative AI models. 

“As part of our work with the rideshare drivers, we [WAO] build tools that people use in the real world,” Calacci said. “There is a smartphone application that is run out of Princeton University and a webtool. I like building these tools. Leveraging some of the infrastructure from ICDS, including the engineering and compute resources for hosting would be useful in this work.” 

Calacci is collaborating with researchers from the University of Colorado Boulder and Princeton University, as well as many grassroots organizations. Research questions are driven by the needs of the communities being directly impacted. 

“I think researchers have an opportunity to have real impact on how data and algorithms are used on a daily basis in people’s lives,” Calacci said. “We have technical expertise that a lot of advocacy groups and community organizations don’t have access to. As researchers we have an opportunity to create general knowledge and develop systems to help communities. To have impact on policies and people’s lives is what excites me the most.” 

Calacci is also teaching a data science course that “prompts students to think critically about the role they play in society and give them the frameworks they need to evaluate how their work will affect people on the job.” 

 

Enrico Casella 

 

Casella focuses on applications of AI and computer vision to monitor animal’s health and productivity. His research enables the detection of early signs of diseases or abnormal growth rates. 

From a young age, Casella was curious about how technology, wearables, smartphones, and different computational systems worked. He earned a doctoral degree in computer science at the University of Kentucky.  

There, he worked with Dr. Simone Silvestri, associate professor in the department of computer science, on a project that fit wearable applications for horse gaits. This project explored using regular smartwatches on humans, or trainers, riding horses to monitor certain gaits (trot, canter, gallop), aimed to help the trainers not overload the horses. 

Prior to working at Penn State, Casella co-authored a paper with Dr. Melissa Cantor, assistant professor of precision dairy science, looking to predicting bovine respiratory disease in group-housed dairy calves using automated feeders and pedometers.  

“My work is very multidisciplinary,” Casella said. “I’ve worked with psychological scientists and electrical engineers for my work on power conservation in smart grids. I used computer vision to monitor dairy calves growth through the use of depth cameras. It is very labor intensive to weigh animals. Applied machine learning and AI are always adopted in my works. Processing heterogeneous data sources and making predictions are only made possible through machine learning and AI algorithms.” 

At Penn State, Casella plans to expand on this research to benefit dairy cows and heat stress issues during warm-weather seasons.  

“A lot of the work I am interested in is using the sensors or wearables to gather information about the status of the animals, but also to predict future health events in advance and start treatment early which has shown to be more effective,” Casella said. “We are using these thermal cameras and environmental information collected at the farm to see if we can predict the cows’ core body temperature and predict heat stress which affects their productivity. If the cow is pregnant, there could be consequences for future generations as well. We could develop several resources that are sustainable to help cool down the cows.” 

Casella also plans to work with turkey hens using three-dimensional (3D) cameras to monitor the turkey’s growth and biometrics to perform uniformity assessment, which are important aspects of productivity and financial returns. The project has its challenges though, as the turkeys’ plumage will create noise in the data. 

“What is fascinating to me is kind of doubling down on the experiences I had growing up, playing with the first iPhone was a process of discovery. Now applying novel technologies and tools to agriculture and animals is a similar type of exploration that I find truly exciting” Casella added. 

Share

Related Posts

News & Events