Project 1 (Ameeta Agrawal):
Project abstract: User-generated social data such as microblogs, discussion posts, news comments, and product reviews, often contain rich expressions of opinions and diverse perspectives on a broad range of topics from health to socio-political movements. Accessing large amounts of diverse data is relatively easy; however, extracting important, relevant, and accurate information that provides a holistic view of the unique and frequent thoughts remains extremely challenging because data is unstructured, noisy, and often redundant. We will leverage the latest advances in large language models to develop and evaluate automatic text summarization models that can generate concise summaries to help us identify salient pieces of information.
- Keywords: (NLP) Natural Language Processing, Large Language Models
- Faculty Mentor: Ameeta Agrawal
- Department: Computer Science
- Community partner(s): TBD
- Tools to be used: NLP Algorithms, Large Language Models
Project 2 (Heejun Chang):
Project abstract: This project examines how climate change and urban development shift water-related ecosystem (WES) services (e.g., water provision, flood and drought regulation, pollutant retention, aquatic habitat) in the Portland metropolitan area [23,41] using the socio-ecological-technological systems (SETS) framework [34]. The SETS framework offers an integrative an comprehensive modeling platform to better understand how specific WES shifts with the disturbance and management (e.g., green infrastructure) of the water cycle in the urban environment and how different populations will be affected by changing WES [14]. This project will address how biophysical changes in the urban environment are directly associated with environmental justice. In collaboration with the City of Portland, Clean Water Services of Washington County, Oregon Department of Environmental Quality, this project co-produces knowledge and information relevant to creating urban resilience in the changing climate. We will use advanced spatial analysis and modeling, machine learning, and system dynamic modeling to better assess the changing WES under different climate change and land development scenarios.
- Keywords: Climate Change, Urban Development, Water-related ecosystems, Environmental Justice.
- Faculty Mentor: Heejun Chang.
- Department: Geography at Portland State University.
- Community partner(s): City of Portland, Clean Water Services of Washington County, Oregon Department of Environmental Quality.
- Tools to be used: (SETS) Socio-ecological-technological systems framework, Spatial Analysis, Modeling, System Dynamic Modeling.
Project 3 (Jay Gopalakrishnan) (Unavailable for Summer 2024):
Project abstract: This undergraduate research project is in collaboration with a local start-up company, Microstructure Engineering (DUNS# 080430947). The company’s goal is to develop affordable computational tools for predicting microstructure evolution and its effect on the properties of metals. Plastic deformation in ductile crystals is caused by the gliding of a collection of dislocations on specific crystallographic slip planes. Physical properties such as strength, ductility and fatigue resistance are thought to be related to the collective motion and interaction of dislocations with other defects in crystalline metals. After the early fundamental work of [29], who gave an incompatibility equation for a dislocation density tensor, progress has not been easy due to difficulties in formulating constitutive laws from atomistic theory and experimentation. One avenue to recent advances is based on simulation of field dislocation mechanics models of [8 ,19, 52] where atomistic information can be built in. Even assuming translational symmetry to restrict to one-dimension (1D), the complex behavior of these nonlinear models are unknown. Thus sacrificing neither originality nor real-world impact, we are led to 1D models, simple enough for an undergraduate student to make a scholarly impact in the field. The student will learn about field dislocation models and work within a python sandbox we will build for algorithmic study of advanced numerical methods for 1D dislocation models.
- Keywords: Field Dislocation Modes, Python, Algorithmic Study, Numerical Methods for 1D Dislocation Models.
- Faculty Mentor: Jay Gopalakrishnan
- Department: Fariborz Maseeh Department of Mathematics + Statistics
- Community partner(s): Microstructure Engineering Co.
- Tools to be used: 1D models, Dislocation Models, Python, Numerical Methods for 1D Dislocation Models.
Project 4 (Samantha Hartzell):
Project abstract: Recent weather events in Portland, OR lead to recorded temperatures of 115 degrees Fahrenheit in July 2021, shattering previous records. The combination of increasing heat and decreasing soil moisture has lead to unprecedented water stress among the city’s urban vegetation [30]. In order to better plan and manage our green infrastructure under these conditions, it is increasingly necessary to understand the impact of water stress on city vegetation [47]. The student will work with data from the Portland Bureau of Environmental Services (BES) to model vegetation water stress on green roofs using process-based models [42]. Modifying typical ecohydrological models to better represent the green roof environment will allow us to simulate the coupled dynamics of vegetation evapotranspiration and soil moisture over time. These methods will take into account the shallow substrate depth and low soil moisture retention characteristic of the green roof environment. The end result will be an assessment of green roof vegetation mortality risk under current and future climate conditions in Portland, OR.
- Keywords: Environment, Weather Events, Green Infrastructure, Water Stress, City Vegetation.
- Faculty Mentor: Samantha Hartzell.
- Department: Civil and Environmental Engineering at Portland State University.
- Community partner(s): (BES) Portland Bureau of Environment Services.
- Tools to be used: Process-Based Models, Modifying Typical Ecohydrological Models.
Project 5 (Sirisha Kothuri):
- Keywords: Transportation Operations, Machine Learning, Requisite Data, Model Formulation and Estimation.
- Faculty Mentor: Sirisha Kothuri.
- Department: Civil and Environmental Engineering.
- Community partner(s): (PORTAL) The regional transportation data archive for the Portland-Vancouver metropolitan region. Bike-Ped PORTAL a national non-motorized traffic count archive.
- Tools to be used: PORTAL, Bike-Ped PORTAL, Data, Models.
Project 6 (Tammy Lee):
Project abstract: Transportation agencies produce a lot of data: signal timing, stop level event
transit data, data from freeway sensors, etc. These data are collected at resolutions anywhere from 1/20th of a second to every 15 min. The PORTAL project is a data lake that archives the Portland, Oregon – Vancouver, Washington metropolitan region’s transportation data. BikePed Portal is the national non-motorized vehicle data lake. Combined, these projects represent more than 30TB of data from a variety of sources. A goal of this project is to evaluate how current and new data sources can be developed into meaningful applications and dashboards to help partnering agencies understand and visualize their data.
- Keywords: Data visualization, data lake
- Faculty: Tammy Lee
- Department: Transportation Research and Education Center (TREC)
- Community partner(s): Southwest Washington Regional Transportation Council, Oregon METRO.
- Tools to be used: PORTAL
Project 7 (John Lipor):
- Keywords: Geothermal Energy Resources, Data, Python, Carbon-Neutral Energy Sources.
- Faculty Mentor: John Lipor.
- Department: Electrical and Computer Engineering at Portland State University.
- Community partner(s): (USGS) United States Geological Survey – Scientists.
- Tools to be used: Python Data Science Packages
Project 8 (Dorcas Ofori-Boateng) (Unavailable for Summer 2024):
Project abstract: Recent times have witnessed growing evidence that the number and magnitude of adverse atmospheric events, especially those affecting costal areas, tend to exhibit an upward trend and will likely continue to increase. Therefore, the first phase of this project is intended to develop topology-based statistical robustness metrics for mulitlayer critical infrastructures that suffer weather/climate-related threats. This, in turn, would allow the simultaneous assessment of the vulnerability of integrated energy resources such as solar, wind and hydropower, molded as multilayer multiplex networks, and the cascading effects of peer failures on transportation, telecommunication and other sectors, at both local and global levels. At the next final phase, this project will extend the utility of the proposed TDA took to investigate network vulnerability in other domains such as, for example, the response of the brain connector under traumas and diseases as well as the dynamics of various modern omics networks under drug intervention – thereby, linking my research on new topology- and geometry-enhanced methods for network organization across power systems to computational biology and other interdisciplinary applications.
- Keywords: Adverse Atmospheric Events, Topology-Based Statistical Metrics, Network Vulnerabilities in Many Sectors.
- Faculty Mentor: Dorcas Ofori-Boateng
- Department: Mathematics + Statistics at Portland State University.
- Community partner(s): TBD
- Tools to be used: Developing topology-based statistical metrics, interdisciplinary applications.
Project 9 (Banafsheh Rekabdar):
Project abstract: This research proposes novel bio-inspired interactive Reinforcement Learning (RL) algorithms enabling simulated healthcare assistive robots/agents to autonomously make optimal decisions and skillfully perform tasks within the patient room. The aim is to design novel bio-inspired RL algorithms based on Spiking Neural Networks (SNNs) to enable simulated assistive robots/agents to make optimal higher decisions within the patient room autonomously. These include making decisions to reach higher-level and lower-level goals. Examples of higher-level goals within the patient room include cleaning duties, food-related activities, and assisting the patients/nurses. Examples of lower-level goals are grasping, reaching, and moving objects. The proposed research will build upon the previous work by PI on using SNNs to learn and early recognition of spatio-temporal patterns [49–51]. For example, in [49–51], SNNs were applied in a unique format to early classify human gestures, which obtained promising results.
- Keywords: RL algorithms, spatio-temporal patterns
- Faculty Mentor: Banafsheh Rekabdar
- Department: Computer Science
- Community partner(s): OHSU
- Tools to be used: TBD
Project 10 (Christof Teuscher):
Project abstract: The goal of this project is to develop socio-technical agent-based models to generate daily activities of runners. To the best of our knowledge, there is currently no computational model of how runners chose routes on their daily runs. There are models for pedestrians and bicyclists [24] that have been used for city for city and infrastructure planning. With over 60 million active runners in the US, we argue that this population should be modeled and included in infrastructure planning, such as street lights, sidewalks, parks with trails, etc. The student will use a state-machine approach to create agent-based behavior models of runners that will complete daily runs in a model of a real city. The OSMnx Python package will be used to model real cities. The agent-based approach will allow to simulate many thousands of different runners and to generate heat maps, not unlike the Strava heat map. Those heat maps can then be used by city infrastructure planners, but also by runners themselves, e.g, to find popular or lonely running routes. We will work with METRO and the City of Portland as a community partner.
- Keywords: Behavior Models, Runners, Python, Heat Maps.
- Faculty Mentor: Christof Teuscher.
- Department: Electrical and Computer Engineering at Portland State University.
- Community partner(s): METRO, City or Portland.
- Tools to be used: Python, Heat Maps, Real City Models, Agent-Based Behavior Models.
Project 11: Liming Wang
Project abstract: The project develops and applies a strategic planning model of transportation and land use to study how policies and technology scenarios affect greenhouse gas emission over the long term. Strategic modeling allows many scenarios to be simulated within a reasonable timeframe, so that stakeholders can quickly go through many different policies and decide where to focus their attention next for more refined studies. The project further develops the VisionEval strategic model emerged from a collaboration between Federal Highway Administration, Oregon DOT, and several other public agencies across the US. We will apply the improved model to evaluate the effect of climate strategies in Oregon. In their role in this project, students will learn to work with strategic simulation models, utilize statistical analysis, geographic information systems (GIS), and data visualization. The research is partnered with Oregon DOT to help them craft their climate strategies.
- Keywords: Data visualization, statistical analysis
- Faculty Mentor: Liming Wang
- Department: School of Urban Studies and Planning
- Community partner(s): Federal Highway Administration, Oregon DOT
- Tools to be used: TBD
References:
Ameeta Agrawal
[6] Question generation by transformers, 2019. arXiv preprint arXiv:1909.05017.
[7] English machine reading comprehension datasets: A survey, 2021. arXiv preprint
arXiv:2101.10421.
Heejun Chang
[23] W. Hoyer and H. Chang. Assessment of freshwater ecosystem services in the Tualatin and
Yamhill basins under climate change and urbanization. Applied Geography, 53:402–416, 2014.
[41] A. Ross and H. Chang. Modeling system dynamics of irrigators’ resilience to cli-
mate change in a glacier-influenced watershed. Hydrological Sciences Journal, 2021. https://www.tandfonline.com/doi/full/10.1080/02626667.2021.1937179.
[34] S. Markolf, M. Chester, D. Eisenberg, D. Iwaniec, B. Ruddell, C. Davidson, R. Zimmerman, T. Miller, and H. Chang. Interdependent infrastructure as linked social, ecological, and techno-logical systems (SETS) to address lock-in and improve resilience. Earth’s Future, 6(12):1638–1659, 2018.
[14] H. Chang, A. Pallathadka, J. Sauer, N. Grimm, R. Zimmerman, C. Cheng, D. Iwaniec, Y. Kim, R. Lloyd, T. McPhearson, B. Rosenzweig, T. Troxler, C. Welty, R. Brenner, and P. Herreros-Cantis. Assessment of urban flood vulnerability using the social-ecological-technological systems framework in six US cities. Sustainable Cities and Society, 68:102786, 2021.
[53] Gelsey, K.* Chang. H. Ramirez, D*. (2023) Effects of landscape characteristics, anthropogenic factors, and seasonality on water quality in Portland, Oregon, Environmental Monitoring and Assessment, 195, 219.
[54] Ramirez, D*, Chang. H., Gelsey, K.* (2022) Effects of Antecedent Precipitation Amount and COVID-19 Lockdown on Water Quality along an Urban Gradient, Hydrology, 9(12), 220.
Jay Gopalakrishnan
[29] Ekkehart Kröner. Allgemeine Kontinuumstheorie der Versetzungen und Eigenspannungen. Arch.
Rational Mech. Anal., 4:273–334 (1960), 1960.
[8] Amit Acharya. A model of crystal plasticity based on the theory of continuously distributed
dislocations. Journal of the Mechanics and Physics of Solids, 49(4):761–784, 2001.
[19] K. Gbemou, V. Taupin, J. M. Raulot, and C. Fressengeas. Building compact dislocation cores
in an elasto-plastic modelof dislocationfields. International Journal of Plasticity, 82:241–259,
2016.
[52] Xiaohan Zhang, Amit Acharya, Noel J. Walkington, and Jacobo Bielak. A single theory for
some quasi-static, supersonic, atomic, andtectonic scale applications of dislocations. Journal of
the Mechanics and Physics of Solids, 84:145–195, 2015.
Samantha Hartzell
[30] K. L. Larson, C. Polsky, P. Gober, H. Chang, and V. Shandas. Vulnerability of water systems to
the effects of climate change and urbanization: A comparison of Phoenix, Arizona and Portland,
Oregon (USA). Environ. Manage, 52:179–195, 2013.
[47] I. Wagner and P. Breil. The role of ecohydrology in creating more resilient cities. Ecohydrol.
Hydrobiol., 13:113–134, 2013.
[42] A. Porporato S. Hartzell, M. S. Bartlett. Unified representation of the C3, C4, and CAM photo-
synthetic pathways with the photo3 model. Ecol. Modell., 384:173–187, 2018.
Sirisha Kothuri
[28] S. Kothuri, K. Nordback, A. Schrope, T. Phillips, and M. Figliozzi. Bicycle and pedestrian
counts at signalized intersections using existing infrastructure: Opportunities and challenges.
Transportation Research Record, pages 11–18, 2017.
[38] K. Nordback, S. Kothuri, T. Phillips, C. Gorecki, and M. Figliozzi. Accuracy of bicycle counting
with pneumatic tubes in oregon. Transportation Research Record, pages 9–17, 2016.
John Lipor
[49] Colin F Williams, Marshall J Reed, Robert H Mariner, Jacob DeAngelo, and S Peter Galanis. Assessment of moderate-and high-temperature geothermal resources of the united states. Tech-
nical report, Geological Survey (US), 2008.
[18] James E Faulds, Stephen Brown, Mark Coolbaugh, Jacob DeAngelo, John H Queen, Sven Tre-itel, Michael Fehler, Eli Mlawsky, Jonathan M Glen, Cary Lindsey, et al. Preliminary report on applications of machine learning techniques to the nevada geothermal play fairway analysis. In Proceedings, 45th Workshop on Geothermal Reservoir Engineering, Stanford University, Stanford, CA, SGP-TR-216, volume 1, 2020.
[10] Jessa Bekker and Jesse Davis. Learning from positive and unlabeled data: A survey. Machine
Learning, 109(4):719–760, 2020.
Banafsheh Rekabdar
[49] Banafsheh Rekabdar, Luke Fraser, Monica Nicolescu, and Mircea Nicolescu. A real-time spike-
timing classifier of spatio-temporal patterns. Neurocomputing, 311:183–196, 2018.
[50] Banafsheh Rekabdar, Monica Nicolescu, Mircea Nicolescu, and Sushil Louis. Using patterns of
firing neurons in spiking neural networks for learning and early recognition of spatio-temporal
patterns. Neural Computing and Applications, 28(5):881–897, 2017.
[51] Banafsheh Rekabdar, Monica Nicolescu, Mircea Nicolescu, Mohammad Taghi Saffar, and Richard Kelley. A scale and translation invariant approach for early classification of spatio-
temporal patterns using spiking neural networks. Neural Processing Letters, 43(2):327–343,2016.
Christof Teuscher
[24] Tetsuro Hyodo, Norikazu Suzuki, and Katsumi Takahashi. Modeling of bicycle route and
destination choice behavior for bicycle road network plan. Transportation Research Record,
1705(1):70–76, 2000.