Campus News
Five student projects funded by 2025-26 CITRIS Tech for Social Good Program
This program supports cross-disciplinary student projects and events aiming to address significant social challenges with technology-based solutions.
The Center for Information Technology Research in the Interest of Society (CITRIS) at UC Santa Cruz has selected five student projects for funding through the 2025-26 CITRIS Tech for Social Good Program. This program supports cross-disciplinary student projects and events aiming to address significant social challenges with technology-based solutions. The program will kick off with initial project team meetings in January. Moving forward, the teams will have monthly check-ins and will present their progress during a project showcase at the end of May of 2026.
All undergraduate, graduate, and postdoctoral students across all campus disciplines and divisions are eligible to participate. Student teams can receive up to $5,000 for research projects and up to $1,000 for events. This year, the program supports the following five student teams:
Task-conditioned prosthetic vision simulator
Over 200 million people worldwide suffer from retinal degenerative diseases, causing permanent blindness. Retinal prostheses electrically stimulate remaining retinal cells to create visual perception. However, current devices like Argus II provide only 60-400 electrodes versus the natural retina’s 1 million cells, creating a severe information bottleneck.
Current prosthetic vision systems use static encoding, treating all visual information equally. A dangerous stove edge receives the same processing as irrelevant wall texture. Users see grainy, unclear images lacking detail for daily tasks like navigating, grasping objects, or reading signs. This is why 60% abandon their devices within a year.
The core challenge is the difficulty of showing what matters the most with only 400 “pixels” available. Current systems waste bandwidth on random information instead of intelligently prioritizing safety hazards, graspable objects, and readable text.
The team will develop a hybrid multi-task prosthetic vision encoder that uses artificial intelligence to automatically understand and prioritize visual information. Instead of blindly converting pixels to electrical signals, this system uses a vision-language model called CLIP that has been trained on 400 million images to understand the semantic meaning. As a result, the users will have clearer, safer, and more useful vision.
The team will implement this system entirely in software using Python, PyTorch (for neural networks), and pulse2percept (for realistic vision simulation). The CLIP encoder remains frozen (pretrained), while we train only small adapter networks on synthetic and real-world images with ground-truth labels for edges, objects, and text regions. Performance will be validated by comparing the system against three baselines: static downsampling, edge-only detection, and single-task optimized encoders.
The project uses scalable cloud infrastructure (Google Colab/AWS) for training and cloud storage (Google Drive/S3) with GitHub version control for reproducibility. This cloud-native approach would allow scaling for greater datasets. The system is designed for edge device deployment like nanos, making it ready for real-time hardware integration and future clinical trials.
Gateways digital literacy jail courses
Music-making serves as an accessible, culturally resonant entry point for students who may not see themselves represented in traditional academic or technical pathways. Digital music tools invite experimentation and emotional expression while building essential digital competencies such as file navigation, editing workflows, and creative problem-solving. This approach creates a learning environment that feels empowering, relevant, and joyful.
Alongside this, the team’s structured design process guides students through planning, prototyping, revising, and completing a final project. It becomes a transferable blueprint that supports follow-through, organization, and intentional project development during incarceration and upon reentry. This directly addresses a major gap in carceral education: students are taught skills, but rarely given a consistent process for building and completing projects.
Gateways will continue using open-source tools such as Blender, ensuring students can access creative technologies post-release without financial barriers. Classes also integrate professional development practices such as portfolio building, interview preparation, and digital identity skills.
By combining creative technology, digital skill-building, music-making, and structured design practices, Gateways aims to reduce the digital divide and strengthen reentry outcomes. The project’s approach emphasizes community, creativity, and long-term empowerment, offering students who are incarcerated not just tools, but pathways, confidence, and sustainable growth beyond release.
Accessibility navigation project
The project aims to create an alternative navigation/map tool for UCSC, centered on accessibility. Currently, campus navigation poses several challenges for members of the campus community with disabilities (e.g., steep inclines, pollen, unpaved routes, uneven roads, etc.). These challenges mean that campus accessibility is limited or the current means of access are laborious (i.e., having to wait for a shuttle). Most students with disabilities often have to rely on shuttle services or navigate the campus with limited mobility options, alienating them from the larger student community.
The team aims to create a navigation tool for students based on their preferences concerning their bodies’ capabilities. The new navigation model will aid the user in finding alternative routes or paths to engage with based on their needs. This will take into account a variety of disabilities, ranging from physical impairments to more transient needs, such as those prone to migraines needing to avoid sunlight. The study will incorporate a research through design model that will engage with various student stakeholders through surveys, interviews, walkthrough studies, codesign workshops, and citizen science models. For the citizen science models, due to the wide number of variables we aim to measure, we require the use of foot sensors to map the wider variables, such as stress points, tactile experiences of using different routes, and the various route alternatives appropriate for the user’s needs.
Slugiculture
Small farmers often find out about crop damage or pest outbreaks after they have grown into a major problem, when they are spotted during walkthroughs. While large farms can afford to use AI and robotic survey data to inform early intervention procedures, small farmers must still deal with these avoidable losses. There is a need for a cost-effective, easy-to-use system which lets farmers discover destructive circumstances before they become major losses, without labor-intensive inspection.
AIRWISE
AIRWISE aims to address the problem of gaps within air quality monitoring networks. To do this, the team will develop a low-cost, solar-powered network of air quality sensor nodes that can operate independently of WiFi. Currently, the best air sensors on the market, PurpleAir sensors, do not meet these criteria and are expensive for low-income communities. The cheapest in-stock sensor is $240 and requires WiFi to upload data. Communities that are impacted by poor air quality and unreliable WiFi cannot easily install these air sensors to obtain proof of their substandard environmental conditions. The only option is to spend thousands of dollars on these sensors and only install them in limited areas.
In addition to these impacted communities, farms are often heavily impacted by poor air quality. In the surrounding areas, pesticides and other gases impact the lives of people living there. WiFi is often unreliable across a large, rural farm, so this air quality data is also difficult to collect. These rural and agricultural zones are exactly where monitoring is needed to protect workers and nearby communities.
Another issue the team aims to address is wildfire detection, which is especially important because many wildfires start in remote areas with no WiFi and then later spread to populated communities. Placing AIRWISE nodes in areas susceptible to wildfires can provide an early warning when a wildfire has started, or even help develop a prediction model to determine whether there is a high risk of wildfire. This can be implemented by setting threshold values for humidity, temperature, and particulate matter (PM) measurements.
One more route the team wants to pursue is mounting these sensor nodes on drones. This would help analyze air quality over large areas of rural land and quickly create detailed maps of where the air quality is the poorest, which would be a difficult task for fixed, ground-based nodes.
The sensor nodes will use LoRa (Long Range) radio waves to communicate with one another. Each node will measure particulate matter (PPM 1, PPM 2.5, and PPM 10) and additional quantities such as humidity, temperature, air pressure, and VOCs. Each node will send all its data in a packet to a router node, which aggregates, processes, and uploads the data. This data will be publicly displayed on a frontend with a live feed of air quality measurements from our sensors. The sensor nodes do not need to be connected to WiFi and can be placed in semi-remote areas with little to no maintenance needed, as their batteries will be charged with an included solar panel. The router node will need to be connected to WiFi and power, but can still be stationed miles away from the sensor nodes and reliably receive data. For LoRa communication, we will use Meshtastic, an open-source community-based mesh network solution for radio communication. This has the additional benefit of increasing our range when other Meshtastic community members set up LoRa nodes that can repeat our messages and serve as intermediaries.
Future funding opportunities through Tech for Social Good Program
For students in search of funding opportunities for their research projects, the application for the 2026-27 Tech for Social Good program will open in mid-November 2026.
The Center for Information Technology Research in the Interest of Society (CITRIS) and the Banatao Institute create information technology solutions for society’s most pressing challenges. Established in 2001, the center leverages the interdisciplinary research strengths of multiple Campuses – Berkeley, Davis, Merced, and Santa Cruz – along with public and private partners to advance the University of California’s mission and the innovative spirit of California. The institute was created to shorten the pipeline between world-class laboratory research and the development of cutting-edge applications, platforms, companies, and even new industries. Find out more at CITRIS.sites.UCSC.edu and CITRIS-UC.org.