Ateneo seeks collaborative trials for machine learning innovation

The Ateneo Laboratory for Intelligent Visual Environments is seeking to co-develop machine learning solutions with industry experts to address complex challenges in medicine, urban planning, and public safety.

The push for interdisciplinary collaboration was the central theme of the second Ateneo Breakthroughs lecture held last 26 February at Escaler Hall. During the event, Dr. Patricia “Pai” Angela R. Abu, an associate professor and chair of the Ateneo de Manila University Department of Information Systems and Computer Science, delivered a talk titled “Smarter Sight: Building Intelligent Visual Systems for Public Good.”

Abu, who leads the ALIVE team, explained that while machine learning allows computers to identify patterns that challenge human experts, these systems struggle with tasks humans find intuitive.

While a toddler can recognize a face with little instruction, computer vision systems require massive datasets and constant testing to handle changes in lighting, angles, and weather.

“Building a reliable machine-learning system requires bridging messy reality and mathematical models,” Abu said, saying that systems must be proven to hold up under real-world conditions rather than just in controlled laboratory demos.

The ALIVE laboratory has already developed several applications aimed at public welfare. In healthcare, the team has produced dental imaging support tools and deep learning models to detect bone metastasis, helping specialists identify subtle patterns at scale.

In urban management, the laboratory developed V-PROBE (Vehicle and Pedestrian Real-Time Observation and Behavioral Evaluation), a system designed to monitor traffic, predict parking availability, and flag congestion.

The laboratory’s current priority is to move research beyond the lab by partnering with industry leaders. Abu cited that these partners provide the operational environments, data pipelines, and domain expertise necessary to test systems against practical constraints such as speed, privacy and hardware limitations.

By working with external stakeholders, research teams can ensure that laboratory experiments translate into innovations that meet the specific needs of end users.