machine learning

Machine Learning Encoder Improves Weather Forecasting and Tsunami Prediction

admin — Wed, 26 Mar 2025 01:20:41 +0000

Machine Learning Encoder Improves Weather Forecasting and Tsunami Prediction admin Tue, 03/25/2025 - 21:20

Successful test results of a new machine learning (ML) technique developed at Georgia Tech could help communities prepare for extreme weather and coastal flooding. The approach could also be applied to other models that predict how natural systems impact society.

Ph.D. student Phillip Si and Assistant Professor Peng Chen developed Latent-EnSF, a technique that improves how ML models assimilate data to make predictions.

In experiments predicting medium-range weather forecasting and shallow water wave propagation, Latent-EnSF demonstrated higher accuracy, faster convergence, and greater efficiency than existing methods for sparse data assimilation.

“We are currently involved in an NSF-funded project aimed at providing real-time information on extreme flooding events in Pinellas County, Florida,” said Si, who studies computational science and engineering (CSE).

“We're actively working on integrating Latent-EnSF into the system, which will facilitate accurate and synchronized modeling of natural disasters. This initiative aims to enhance community preparedness and safety measures in response to flooding risks.”

Latent-EnSF outperformed three comparable models in assimilation speed, accuracy, and efficiency in shallow water wave propagation experiments. These tests show models can make better and faster predictions of coastal flood waves, tides, and tsunamis.

In experiments on medium-range weather forecasting, Latent-EnSF surpassed the same three control models in accuracy, convergence, and time. Additionally, this test demonstrated Latent-EnSF's scalability compared to other methods.

These promising results support using ML models to simulate climate, weather, and other complex systems.

Traditionally, such studies require employment of large, energy-intensive supercomputers. However, advances like Latent-EnSF are making smaller, more efficient ML models feasible for these purposes.

The Georgia Tech team mentioned this comparison in its paper. It takes hours for the European Center for Medium-Range Weather Forecasts computer to run its simulations. Conversely, the ML model FourCastNet calculated the same forecast in seconds.

“Resolution, complexity, and data-diversity will continue to increase into the future,” said Chen, an assistant professor in the School of CSE.

“To keep pace with this trend, we believe that ML models and ML-based data assimilation methods will become indispensable for studying large-scale complex systems.”

Data assimilation is the process by which models continuously ingest new, real-world data to update predictions. This data is often sparse, meaning it is limited, incomplete, or unevenly distributed over time.

Latent-EnSF builds on the Ensemble Filter Scores (EnSF) model developed by Florida State University and Oak Ridge National Laboratory researchers.

EnSF’s strength is that it assimilates data with many features and unpredictable relationships between data points. However, integrating sparse data leads to lost information and knowledge gaps in the model. Also, such large models may stop learning entirely from small amounts of sparse data.

The Georgia Tech researchers employ two variational autoencoders (VAEs) in Latent-EnSF to help ML models integrate and use real-world data. The VAEs encode sparse data and predictive models together in the same space to assimilate data more accurately and efficiently.

Integrating models with new methods, like Latent-EnSF, accelerates data assimilation. Producing accurate predictions more quickly during real-world crises could save lives and property for communities.

To share Latent-EnSF to the broader research community, Chen and Si presented their paper at the SIAM Conference on Computational Science and Engineering (CSE25). The Society of Industrial and Applied Mathematics (SIAM) organized CSE25, held March 3-7 in Fort Worth, Texas.

Chen was one of ten School of CSE faculty members who presented research at CSE25, representing one-third of the School’s faculty body. Latent-EnSF was one of 15 papers by School of CSE authors and one of 23 Georgia Tech papers presented at the conference.

The pair will also present Latent-EnSF at the upcoming International Conference on Learning Representations (ICLR 2025). Occurring April 24-28 in Singapore, ICLR is one of the world’s most prestigious conferences dedicated to artificial intelligence research.

“We hope to bring attention to experts and domain scientists the exciting area of ML-based data assimilation by presenting our paper,” Chen said. “Our work offers a new solution to address some of the key shortcomings in the area for broader applications.”

Summary sentence

Ph.D. student Phillip Si and Assistant Professor Peng Chen developed Latent-EnSF, a technique that improves how ML models assimilate data to make predictions.

Summary

Ph.D. student Phillip Si and Assistant Professor Peng Chen developed Latent-EnSF, a technique that improves how ML models assimilate data to make predictions.

Dateline

Fri, 03/14/2025 - 12:00

Contact

Bryant Wine, Communications Officer
bryant.wine@cc.gatech.edu

Keywords

College of Computing

School of Computational Science and Engineering

go-researchnews

Research Horizons

Daily Digest

Georgia Tech News Center

machine learning

artificial intelligence

News room topics

Earth and Environment

Science and Technology

Categories

City Planning, Transportation, and Urban Growth

Community

Computer Science/Information Technology and Security

Engineering

Environment

Research

Student and Faculty

Student Research

Mercury ID

681164

Source updated

Tue, 03/25/2025 - 21:19

Faculty to use AI for Protein Design and Discovery with Support of $1.8 Million NIH Grant

admin — Mon, 13 May 2024 14:38:41 +0000

Faculty to use AI for Protein Design and Discovery with Support of $1.8 Million NIH Grant admin Mon, 05/13/2024 - 10:38

The National Institute of Health (NIH) has awarded Yunan Luo a grant for more than $1.8 million to use artificial intelligence (AI) to advance protein research.

New AI models produced through the grant will lead to new methods for the design and discovery of functional proteins. This could yield novel drugs and vaccines, personalized treatments against diseases, and other advances in biomedicine.

“This project provides a new paradigm to analyze proteins’ sequence-structure-function relationships using machine learning approaches,” said Luo, an assistant professor in Georgia Tech’s School of Computational Science and Engineering (CSE).

“We will develop new, ready-to-use computational models for domain scientists, like biologists and chemists. They can use our machine learning tools to guide scientific discovery in their research.”

Luo’s proposal improves on datasets spearheaded by AlphaFold and other recent breakthroughs. His AI algorithms would integrate these datasets and craft new models for practical application.

One of Luo’s goals is to develop machine learning methods that learn statistical representations from the data. This reveals relationships between proteins’ sequence, structure, and function. Scientists then could characterize how sequence and structure determine the function of a protein.

Next, Luo wants to make accurate and interpretable predictions about protein functions. His plan is to create biology-informed deep learning frameworks. These frameworks could make predictions about a protein’s function from knowledge of its sequence and structure. It can also account for variables like mutations.

In the end, Luo would have the data and tools to assist in the discovery of functional proteins. He will use these to build a computational platform of AI models, algorithms, and frameworks that ‘invent’ proteins. The platform figures the sequence and structure necessary to achieve a designed proteins desired functions and characteristics.

“My students play a very important part in this research because they are the driving force behind various aspects of this project at the intersection of computational science and protein biology,” Luo said.

“I think this project provides a unique opportunity to train our students in CSE to learn the real-world challenges facing scientific and engineering problems, and how to integrate computational methods to solve those problems.”

The $1.8 million grant is funded through the Maximizing Investigators’ Research Award (MIRA). The National Institute of General Medical Sciences (NIGMS) manages the MIRA program. NIGMS is one of 27 institutes and centers under NIH.

MIRA is oriented toward launching the research endeavors of young career faculty. The grant provides researchers with more stability and flexibility through five years of funding. This enhances scientific productivity and improves the chances for important breakthroughs.

Luo becomes the second School of CSE faculty to receive the MIRA grant. NIH awarded the grant to Xiuwei Zhang in 2021. Zhang is the J.Z. Liang Early-Career Assistant Professor in the School of CSE.

“After NIH, of course, I first thanked my students because they laid the groundwork for what we seek to achieve in our grant proposal,” said Luo.

“I would like to thank my colleague, Xiuwei Zhang, for her mentorship in preparing the proposal. I also thank our school chair, Haesun Park, for her help and support while starting my career.”

Summary sentence

The National Institute of Health (NIH) has awarded Yunan Luo a grant for more than $1.8 million to use artificial intelligence (AI) to advance protein research.

Summary

The National Institute of Health (NIH) has awarded Yunan Luo a grant for more than $1.8 million to use artificial intelligence (AI) to advance protein research.

Dateline

Wed, 11/29/2023 - 12:00

Contact

Bryant Wine, Communications Officer
bryant.wine@cc.gatech.edu

Keywords

go-researchnews

Daily Digest

College of Computing

School of Computational Science and Engineering

Computational Science and Engineering

artificial intelligence

machine learning

go-ai

Research Horizons

News room topics

Science and Technology

Categories

Biotechnology, Health, Bioengineering, Genetics

Computer Science/Information Technology and Security

Life Sciences and Biology

Research

Student and Faculty

Mercury ID

671293

Source updated

Mon, 05/13/2024 - 10:37

Machine Learning Could be Key to Early Leakage Detection in Underground Carbon Storage Sites

admin — Mon, 13 May 2024 14:38:41 +0000

Machine Learning Could be Key to Early Leakage Detection in Underground Carbon Storage Sites admin Mon, 05/13/2024 - 10:38

A new machine learning method could help engineers detect leaks in underground reservoirs earlier, mitigating risks associated with geological carbon storage (GCS). Further study could advance machine learning capabilities while improving safety and efficiency of GCS.

The feasibility study by Georgia Tech researchers explores using conditional normalizing flows (CNFs) to convert seismic data points into usable information and observable images. This potential ability could make monitoring underground storage sites more practical and studying the behavior of carbon dioxide plumes easier.

The 2023 Conference on Neural Information Processing Systems (NeurIPS 2023) accepted the group’s paper for presentation. They presented their study on Dec. 16 at the conference’s workshop on Tackling Climate Change with Machine Learning.

“One area where our group excels is that we care about realism in our simulations,” said Professor Felix Herrmann. “We worked on a real-sized setting with the complexities one would experience when working in real-life scenarios to understand the dynamics of carbon dioxide plumes.”

CNFs are generative models that use data to produce images. They can also fill in the blanks by making predictions to complete an image despite missing or noisy data. This functionality is ideal for this application because data streaming from GCS reservoirs are often noisy, meaning it’s incomplete, outdated, or unstructured data.

The group found in 36 test samples that CNFs could infer scenarios with and without leakage using seismic data. In simulations with leakage, the models generated images that were 96% similar to ground truths. CNFs further supported this by producing images 97% comparable to ground truths in cases with no leakage.

This CNF-based method also improves current techniques that struggle to provide accurate information on the spatial extent of leakage. Conditioning CNFs to samples that change over time allows it to describe and predict the behavior of carbon dioxide plumes.

This study is part of the group’s broader effort to produce digital twins for seismic monitoring of underground storage. A digital twin is a virtual model of a physical object. Digital twins are commonplace in manufacturing, healthcare, environmental monitoring, and other industries.

“There are very few digital twins in earth sciences, especially based on machine learning,” Herrmann explained. “This paper is just a prelude to building an uncertainty aware digital twin for geological carbon storage.”

Herrmann holds joint appointments in the Schools of Earth and Atmospheric Sciences (EAS), Electrical and Computer Engineering, and Computational Science and Engineering (CSE).

School of EAS Ph.D. student Abhinov Prakash Gahlot is the paper’s first author. Ting-Ying (Rosen) Yu (B.S. ECE 2023) started the research as an undergraduate group member. School of CSE Ph.D. students Huseyin Tuna Erdinc, Rafael Orozco, and Ziyi (Francis) Yin co-authored with Gahlot and Herrmann.

NeurIPS 2023 took place Dec. 10-16 in New Orleans. Occurring annually, it is one of the largest conferences in the world dedicated to machine learning.

Over 130 Georgia Tech researchers presented more than 60 papers and posters at NeurIPS 2023. One-third of CSE’s faculty represented the School at the conference. Along with Herrmann, these faculty included Ümit Çatalyürek, Polo Chau, Bo Dai, Srijan Kumar, Yunan Luo, Anqi Wu, and Chao Zhang.

“In the field of geophysics, inverse problems and statistical solutions of these problems are known, but no one has been able to characterize these statistics in a realistic way,” Herrmann said.

“That’s where these machine learning techniques come into play, and we can do things now that you could never do before.”

Summary sentence

Summary

Dateline

Wed, 12/20/2023 - 12:00

Contact

Bryant Wine, Communications Officer
bryant.wine@cc.gatech.edu

Associated importer

Keywords

go-researchnews

Daily Digest

College of Computing

School of Computational Science and Engineering

Computational Science and Engineering

artificial intelligence

machine learning

College of Sciences

School of Earth and Atmospheric Sciences

college of engineering

College of Engineering; School of Electrical and Computer Engineering

go-ai

Research Horizons

News room topics

Earth and Environment

Categories

Computer Science/Information Technology and Security

Mercury ID

671661

Source updated

Mon, 05/13/2024 - 10:37

Researchers Leverage AI to Develop Early Diagnostic Test for Ovarian Cancer

admin — Tue, 30 Jan 2024 15:55:40 +0000

Researchers Leverage AI to Develop Early Diagnostic Test for Ovarian Cancer admin Tue, 01/30/2024 - 10:55

For over three decades, a highly accurate early diagnostic test for ovarian cancer has eluded physicians. Now, scientists in the Georgia Tech Integrated Cancer Research Center (ICRC) have combined machine learning with information on blood metabolites to develop a new test able to detect ovarian cancer with 93 percent accuracy among samples from the team’s study group.

John McDonald, professor emeritus in the School of Biological Sciences, founding director of the ICRC, and the study’s corresponding author, explains that the new test’s accuracy is better in detecting ovarian cancer than existing tests for women clinically classified as normal, with a particular improvement in detecting early-stage ovarian disease in that cohort.

The team’s results and methodologies are detailed in a new paper, “A Personalized Probabilistic Approach to Ovarian Cancer Diagnostics,” published in the March 2024 online issue of the medical journal Gynecologic Oncology. Based on their computer models, the researchers have developed what they believe will be a more clinically useful approach to ovarian cancer diagnosis — whereby a patient’s individual metabolic profile can be used to assign a more accurate probability of the presence or absence of the disease.

“This personalized, probabilistic approach to cancer diagnostics is more clinically informative and accurate than traditional binary (yes/no) tests,” McDonald says. “It represents a promising new direction in the early detection of ovarian cancer, and perhaps other cancers as well.”

The study co-authors also include Dongjo Ban, a Bioinformatics Ph.D. student in McDonald’s lab; Research Scientists Stephen N. Housley, Lilya V. Matyunina, and L.DeEtte (Walker) McDonald; Regents’ Professor Jeffrey Skolnick, who also serves as Mary and Maisie Gibson Chair in the School of Biological Sciences and Georgia Research Alliance Eminent Scholar in Computational Systems Biology; and two collaborating physicians: University of North Carolina Professor Victoria L. Bae-Jump and Ovarian Cancer Institute of Atlanta Founder and Chief Executive Officer Benedict B. Benigno. Members of the research team are forming a startup to transfer and commercialize the technology, and plan to seek requisite trials and FDA approval for the test.

Silent killer

Ovarian cancer is often referred to as the silent killer because the disease is typically asymptomatic when it first arises — and is usually not detected until later stages of development, when it is difficult to treat.

McDonald explains that while the average five-year survival rate for late-stage ovarian cancer patients, even after treatment, is around 31 percent — but that if ovarian cancer is detected and treated early, the average five-year survival rate is more than 90 percent.

“Clearly, there is a tremendous need for an accurate early diagnostic test for this insidious disease,” McDonald says.

And although development of an early detection test for ovarian cancer has been vigorously pursued for more than three decades, the development of early, accurate diagnostic tests has proven elusive. Because cancer begins on the molecular level, McDonald explains, there are multiple possible pathways capable of leading to even the same cancer type.

“Because of this high-level molecular heterogeneity among patients, the identification of a single universal diagnostic biomarker of ovarian cancer has not been possible,” McDonald says. “For this reason, we opted to use a branch of artificial intelligence — machine learning — to develop an alternative probabilistic approach to the challenge of ovarian cancer diagnostics.”

Metabolic profiles

Georgia Tech co-author Dongjo Ban, whose thesis research contributed to the study, explains that “because end-point changes on the metabolic level are known to be reflective of underlying changes operating collectively on multiple molecular levels, we chose metabolic profiles as the backbone of our analysis.”

“The set of human metabolites is a collective measure of the health of cells,” adds coauthor Jeffrey Skolnick, “and by not arbitrarily choosing any subset in advance, one lets the artificial intelligence figure out which are the key players for a given individual.”

Mass spectrometry can identify the presence of metabolites in the blood by detecting their mass and charge signatures. However, Ban says, the precise chemical makeup of a metabolite requires much more extensive characterization.

Ban explains that because the precise chemical composition of less than seven percent of the metabolites circulating in human blood have, thus far, been chemically characterized, it is currently impossible to accurately pinpoint the specific molecular processes contributing to an individual's metabolic profile.

However, the research team recognized that, even without knowing the precise chemical make-up of each individual metabolite, the mere presence of different metabolites in the blood of different individuals, as detected by mass spectrometry, can be incorporated as features in the building of accurate machine learning-based predictive models (similar to the use of individual facial features in the building of facial pattern recognition algorithms).

“Thousands of metabolites are known to be circulating in the human bloodstream, and they can be readily and accurately detected by mass spectrometry and combined with machine learning to establish an accurate ovarian cancer diagnostic,” Ban says.

A new probabilistic approach

The researchers developed their integrative approach by combining metabolomic profiles and machine learning-based classifiers to establish a diagnostic test with 93 percent accuracy when tested on 564 women from Georgia, North Carolina, Philadelphia and Western Canada. 431 of the study participants were active ovarian cancer patients, and while the remaining 133 women in the study did not have ovarian cancer.

Further studies have been initiated to study the possibility that the test is able to detect very early-stage disease in women displaying no clinical symptoms, McDonald says.

McDonald anticipates a clinical future where a person with a metabolic profile that falls within a score range that makes cancer highly unlikely would only require yearly monitoring. But someone with a metabolic score that lies in a range where a majority (say, 90%) have previously been diagnosed with ovarian cancer would likely be monitored more frequently — or perhaps immediately referred for advanced screening.

Citation: https://doi.org/10.1016/j.ygyno.2023.12.030

Funding

This research was funded by the Ovarian Cancer Institute (Atlanta), the Laura Crandall Brown Foundation, the Deborah Nash Endowment Fund, Northside Hospital (Atlanta), and the Mark Light Integrated Cancer Research Student Fellowship.

Disclosure

Study co-authors John McDonald, Stephen N. Housley, Jeffrey Skolnick, and Benedict B. Benigno are the co-founders of MyOncoDx, Inc., formed to support further research, technology transfer, and commercialization for the team’s new clinical tool for the diagnosis of ovarian cancer.

Subtitle

The Georgia Tech Integrated Cancer Research Center has combined machine learning with information on blood metabolites to develop a new early diagnostic test that detects ovarian cancer with 93 percent accuracy.

Summary sentence

Summary

Dateline

Mon, 01/29/2024 - 12:00

renay.san@cos.gatech.edu

Contact

Writer: Renay San Miguel
Communications Officer II/Science Writer
College of Sciences
404-894-5209

Editor: Jess Hunt-Ralston

Keywords

College of Sciences

School of Biological Sciences

John McDonald

Dongio Ban

Jeffrey Skolnick

Stephen N. Housley

Lilya Matyunina

LeDette Walker McDonald

ovarian cancer

Ovarian Cancer Institute

Benedict Benigno

diagnostic tests

machine learning

cos-microbial

cos-research

go-bio

go-ai

go-researchnews

News room topics

Health and Medicine

Categories

Cancer Research

Life Sciences and Biology

Research

Student and Faculty

Mercury ID

672550

Source updated

Tue, 01/30/2024 - 10:54

Coskun Receives $1.86 Million NIH MIRA Award to Map Spatial Molecular Neighborhoods

admin — Wed, 08 Nov 2023 17:24:27 +0000

Coskun Receives $1.86 Million NIH MIRA Award to Map Spatial Molecular Neighborhoods admin Wed, 11/08/2023 - 12:24

Ahmet Coskun has a saying on the homepage of his lab’s website: “Seeing is believing. Quantifying is proving.” So, with that in mind, Coskun and his team have developed multiplex imaging tools and combined them with machine learning techniques – for believing and quantifying.

Now, to support Coskun’s research, the National Institutes of Health has granted him the prestigious Maximizing Investigator’s Research Award (MIRA) from the National Institute of General Medical Sciences. Coskun and his team will use the five-year, $1.86 million award for a project entitled, “Dissecting subcellular and cellular organization by spatial molecular neighborhood networks.”

They plan to probe subcellular and cellular organization, counting molecular neighborhoods and building maps to help researchers better understand the spatial organization of cells and molecules, insights that can open the door to game-changing personalized treatments for multiple diseases.

“The spatial organization of these neighborhoods, of RNA and protein molecules, is important for cellular function,” said Coskun, a Bernie Marcus Early Career Professor in the Coulter Department of Biomedical Engineering at Georgia Tech and Emory University. “So, we’re basically making maps of molecules within the cell.”

The maps can ultimately help researchers identify cell types that would best treat various diseases, while also explaining why some patients will respond to a particular treatment, and others won’t.

The NIH’s MIRA program provides researchers with greater stability and flexibility in funding while enhancing their ability to creatively tackle ambitious scientific problems. And part of the aim, said Coskun, “is to address basic biology questions that have implications on multiple diseases in the future. This single cell work has that kind of potential.”

For Coskun, the MIRA is the next phase of support in a flurry of awards that have come his way recently: it’s the fifth NIH grant his lab has received this year, with a total value of $3.6 million.

“This year has been a great year for us,” said Coskun. “It’s encouraging to receive this kind of recognition and support for research and technology that we believe will play an important role in the lives of patients.”

Subtitle

Summary sentence

Georgia Tech researcher plans to probe subcellular and cellular organization, counting molecular neighborhoods and building to better understand the spatial organization of cells.

Summary

Georgia Tech researcher probing subcellular and cellular organization, counting molecular neighborhoods and building maps to better understand the spatial organization of cells and molecules, opening the door to game-changing personalized treatments for multiple diseases.

Dateline

Wed, 11/08/2023 - 12:00

jerry.grillo@ibb.gatech.edu

Contact

Writer: Jerry Grillo

Associated importer

Keywords

multiplexed imaging

go-bio

machine learning

National Institutes of Health

Ahmet Coskun

News room topics

Health and Medicine

Science and Technology

Categories

Biotechnology, Health, Bioengineering, Genetics

Mercury ID

670950

Source updated

Wed, 11/08/2023 - 12:24

'Human Rights' May Help Shape Artificial Intelligence in 2019

admin — Tue, 06 Jun 2023 17:15:18 +0000

'Human Rights' May Help Shape Artificial Intelligence in 2019 admin Tue, 06/06/2023 - 13:15

Ethics and accountability will be among the most significant challenges for artificial intelligence (AI) in 2019, according to a survey of researchers at Georgia Tech’s College of Computing.

In response to an email query about AI developments that can be expected in 2019, most of the researchers – whether talking about machine learning (ML), robotics, data visualizations, natural language processing, or other facets of AI – touched on the growing importance of recognizing the needs of people in AI systems.

“In 2019, I hope we will see AI researchers and practitioners start to frame the debate about proper and improper uses of artificial intelligence and machine learning in terms of human rights,” said Associate Professor Mark Riedl.

[RELATED: Is AI Coming For My Job?]

“More and more, interpretability and fairness are being recognized as critical issues to address to ensure AI appropriately interacts with society,” said Ph.D. student Fred Hohman.

Taking on algorithmic bias

Questions about the rights of end users of AI-enabled services and products are becoming a priority, but Riedl said more is needed.

“Companies are making progress in recognizing that AI systems may be biased in prejudicial ways. [However,] we need to start talking about the next step: remedy. How do people seek remedy if they believe an AI system made a wrong decision?” said Riedl.

Assistant Professor Jamie Morgenstern sees algorithmic bias as an ongoing concern in 2019 and gave banking as an example of an industry that may be in the news for its algorithmic decision-making.

“I project that we’ll have more high-profile examples of financial systems that use machine learning having worse rates of lending to women, people of color, and other communities historically underrepresented in the ‘standard’ American economic system,” Morgenstern said.

In recent years corporate responses to cases of bias have been hit or miss, but Assistant Professor Munmun De Choudhury said 2019 may see a shift in how tech companies balance their shareholders’ interests with the interests of their customers and society.

“[Companies] will be increasingly subject to governmental regulation and will be forced to come up with safeguards to address misuse and abuse of their technologies, and will even consider broader partnerships with their market competitors to achieve this. For some corporations, business interests may take a backseat to ethics until they regain customer trust,” said De Choudhury.

Working toward more transparency

One way companies can regain that trust is through sharing their algorithms with the public, our experts said.

“Developers tend to walk around feeling objective because ‘it’s the algorithm that is determining the answer’. Moving forward, I believe that the algorithms will have to be increasingly ‘inspectable’ and developers will have to explain their answers,” Executive Associate Dean and Professor Charles Isbell.

Ph.D. student Yuval Pinter agreed. In the coming year, “[I] think we will see that researchers are trying to [develop] techniques and tests that can help us to better understand what’s going on in the actual wiring of our very fancy machine learning models.

“This is not only for curiosity but also because legal applications or regulation in various countries are starting to require that algorithmic decision-making programs be able to explain why they are doing what they are doing,” said Pinter.

Regents’ Professor Ron Arkin believes that these concerns are becoming more central precisely because artificial intelligence will continue to grow in importance in our everyday lives.

[RELATED: Who's Behind the Wheel?]

“Despite continued hype and omnipresent doomsayers, panic and fear over the growth of AI and robotics should begin to subside in 2019 as the benefits to people’s lives are becoming more apparent to the world.

“However, I expect to see lawyers jumping into the fray so we may also see lawsuits determining policy for self-driving cars [and other applications] more so than government regulation or the legal system,” said Arkin.

Subtitle

Georgia Tech experts highlight need to address bias and transparency in ongoing debate about role of AI

Summary sentence

Georgia Tech researchers say ethics and transparency are likely top 2019 trends in the burgeoning field of AI.

Dateline

Tue, 01/15/2019 - 12:00

Contact

Albert Snedeker, Communications Manager

albert.snedeker@cc.gatech.edu

Associated importer

Keywords

artificial intelligence

machine learning

algorithmic bias

transparency

riedl

hohman

isbell

de choudhury

morgenstern

arkin

2019 trends

Mercury ID

616279

Source updated

Fri, 01/25/2019 - 10:27

'Human Rights' May Help Shape Artificial Intelligence in 2019

admin — Tue, 06 Jun 2023 17:15:10 +0000

'Human Rights' May Help Shape Artificial Intelligence in 2019 admin Tue, 06/06/2023 - 13:15

Ethics and accountability will be among the most significant challenges for artificial intelligence (AI) in 2019, according to a survey of researchers at Georgia Tech’s College of Computing.

[RELATED: Is AI Coming For My Job?]

“More and more, interpretability and fairness are being recognized as critical issues to address to ensure AI appropriately interacts with society,” said Ph.D. student Fred Hohman.

Taking on algorithmic bias

Questions about the rights of end users of AI-enabled services and products are becoming a priority, but Riedl said more is needed.

Working toward more transparency

One way companies can regain that trust is through sharing their algorithms with the public, our experts said.

Regents’ Professor Ron Arkin believes that these concerns are becoming more central precisely because artificial intelligence will continue to grow in importance in our everyday lives.

[RELATED: Who's Behind the Wheel?]

Subtitle

Georgia Tech experts highlight need to address bias and transparency in ongoing debate about role of AI

Summary sentence

Georgia Tech researchers say ethics and transparency are likely top 2019 trends in the burgeoning field of AI.

Dateline

Tue, 01/15/2019 - 12:00

Contact

Albert Snedeker, Communications Manager

albert.snedeker@cc.gatech.edu

Associated importer

Keywords

artificial intelligence

machine learning

algorithmic bias

transparency

riedl

hohman

isbell

de choudhury

morgenstern

arkin

2019 trends

Mercury ID

616279

Source updated

Fri, 01/25/2019 - 10:27

Georgia Tech Plans Tokyo Redesign Using Social Data – Including Tweets

bwaye3 — Tue, 19 Apr 2022 16:42:57 +0000

Georgia Tech Plans Tokyo Redesign Using Social Data – Including Tweets bwaye3 Tue, 04/19/2022 - 12:42

What if everyone in a neighborhood had a voice in redesigning it? How can city planners democratize future smart cities, big data analytics and decision-making by involving citizen participation?

That’s the vision of the Tokyo Smart City Studio, using GPS and other location data to track trends in human behavior. The studio is part of the Eco Urban Lab, directed by Perry Yang. In the Studio, Georgia Tech designers are changing the way we plan cities.

According to Yang, designers need to know, “how urban form accommodates flows of all kinds, including information, energy, and human movements, and how patterns of flows are mapped through physical configurations that change over time.” Using advanced analytics on a massive pool of data, designers can see how people choose to use a space. Then, they can make decisions based on real behavior rather than theoretical models or small surveys, Yang said.

Projects from the Studio recently drew Esri’s attention for their advanced use of geographic analytics. As part of the studio, students used large-scale location data to model traffic flows in different segments of Tokyo.

Esri is best known as the company that developed GIS (geographic information system) software which is widely used by city and regional planners. Esri's blog covers innovative use of geospatial data.

To date, the Studio has delivered smart-city proposals for three different Tokyo neighborhoods: Urawa-Misono in 2017, Kyojima in 2018-2019, and Shinagawa in 2020. Yang said this year’s project will involve the Nihonbashi district, home of the Tokyo Stock Exchange and a historic origin of Tokyo’s modern urban culture.

Tokyo is one of the busiest, most populous cities in the world, and as such, it’s difficult to create sustainable, healthy environments, said Yang. His studio focuses on global collaboration to solve the problems facing cities like Tokyo.

For the Nihonbashi district, the Studio is collaborating with the Department of Urban Engineering at the University of Tokyo; the Global Carbon Project; the National Institute for Environmental Studies, Japan; and the Mitsui Fudosan University of Tokyo Laboratory.

Thanks to the diversity of ideas that this collaboration generates, the Tokyo Smart City Studio proposals aren’t purely theoretical. Students work on real neighborhood needs identified by the local government, using real on-site data gathered and processed by groundbreaking methods. Massive amounts of data require AI techniques and algorithms to process some of them.

Design Makes Cities Smart

Unlike a smartphone or a smart home, a smart city is measured by more than its Internet connection or its automation. A city is smart because its design is based on data about flows of people, resources, energy, and information. “Empowered by new tools and technologies, cities are now far more designable than ever before. Future urban design and development are to be data-driven," Yang said.

"Cities are becoming more interactive and situation-driven and have to be more responsive, adaptable, and resilient to future conditions.”

Only in the past few years have smartphones, GPS, and Internet-of-Things (IoT) devices made real data on a city-wide scale available. These new technologies may enable good planning that is driven by social, institutional, and physical contexts, Yang said.

Students Predict Mobility Using AI Techniques

But new data and a new design philosophy require new methods of design, which haven’t been created yet. To meet this challenge, Yang’s students use AI techniques to create models of behavior.

An early test in Urawa-Misono used Tokyo traffic data. The model was “trained” using existing traffic data. Designers then check the model to see if correctly predicts congestion points. If it works, that suggests the model can be used to estimate the impact of design changes on traffic flow.

Methods like this allow designers to move mobility analytics into the design phase. Mobility is key in the urban environment, but mobility studies usually aren’t done until we envision how future urban spaces are shaped through design.

“Mobility research during these studios developed test processes to better integrate mobility metrics with design decisions,” Yang said.

The following studio in Kyojima expanded the role of deep learning techniques. Georgia Tech students participated in street experiments in Tokyo by carrying vital sensors to measure data such as heart beat, temperature, humidity, UV, pressure and acceleration of human movement, to understand human perception and their connections to urban street environments. Students created a rapid automatic assessment of sequential urban experiences: basically, how pleasant a person would experience walking down a street.

The resulting article, Computing Sequential Experience of an Urban Street using Deep Learning Technique, led by Helen Chen, a Ph.D. student in the School of City and Regional Planning, was presented at the 2019 International Conference of Urban Informatics.

Both the Urawa-Misono and Kyojima studio projects made mobility analysis and feedback part of the design phase, yielding designs better supporting resilience, sustainability, and community health.

Using Social Media to Check Risk of Heat Stress

In Tokyo, studio planners expanded mobility studies to include public health. Tokyo officials were concerned that heat waves would pose a risk to the 2020 Olympic crowds.

To address this, researchers used a combination of GPS and social networking service information to track people’s motion and response to heat “in near real-time [and] in high spatial resolution,” said Yang.

Students used thermal images from a helicopter to find hot zones. GPS data showed pedestrian exposure to those hot zones. The combination of heat and pedestrian exposure allowed prediction of risky areas.

In collaboration with partners in Japan, researchers searched geotagged twitter data for terms related to heat discomfort. They found a high correlation between the predicted risk zones and areas where people complained about heat.

Both the prediction of heat risk and the use of social media to check the prediction are new methods developed in the Studio.

Summary sentence

Students use big data and machine learning for neighborhood proposals.

Summary

Students in the Tokyo Smart City studio, of the Eco Urban Lab, are changing the way we plan cities. Using machine learning techniques on data from smartphones, GPS, and Internet-of-Things (IoT) devices, students develop proposals based on measured human experience.

Dateline

Tue, 04/05/2022 - 12:00

ann.hoevel@design.gatech.edu

Location

Atlanta, GA

Associated importer

Keywords

tokyo

Tokyo Smart City studio

Perry Pei-Ju Yang

eco urban lab

machine learning

gis

social media

big data analytics

Internet of Things

go-researchnews

News room topics

Society and Culture

Categories

Student Research

Building Construction

City Planning, Transportation, and Urban Growth

Core research areas

Data Engineering and Science

People and Technology

Public Service, Leadership, and Policy

Mercury ID

657025

Source updated

Thu, 04/07/2022 - 15:05

machine learning

Machine Learning Encoder Improves Weather Forecasting and Tsunami Prediction

Related links

Keywords

News room topics

Categories

Faculty to use AI for Protein Design and Discovery with Support of $1.8 Million NIH Grant

Related links

Keywords

News room topics

Categories

Machine Learning Could be Key to Early Leakage Detection in Underground Carbon Storage Sites

Keywords

News room topics

Categories

Researchers Leverage AI to Develop Early Diagnostic Test for Ovarian Cancer

Related links

Keywords

News room topics

Categories

Coskun Receives $1.86 Million NIH MIRA Award to Map Spatial Molecular Neighborhoods

Keywords

News room topics

Categories

'Human Rights' May Help Shape Artificial Intelligence in 2019

Taking on algorithmic bias

Working toward more transparency

Keywords

'Human Rights' May Help Shape Artificial Intelligence in 2019

Taking on algorithmic bias

Working toward more transparency

Keywords

Georgia Tech Plans Tokyo Redesign Using Social Data – Including Tweets

Design Makes Cities Smart

Students Predict Mobility Using AI Techniques

Using Social Media to Check Risk of Heat Stress

Keywords

News room topics

Categories

Core research areas