MIA - Nadia Irwanto

Nadia was a recipient of our 2018 Summer Grant. However, we have not heard from her since she received the Grant. Nadia, if you read this, please reach out to us.

Nathan Lin - Summer Grant 2018 Recipient

Computer graphics has many applications in preserving cultural heritage. Today, technology can be used to scan artifacts, create 3D models, and much more. The “Reconstruction and Lighting of Assyrian Palace Reliefs” project is focused on creating a scene for scholarly research or museum exhibition containing 3D models of Assyrian stone reliefs. 

 

For this particular scene, 3D models were created of two stone reliefs from the Northwest Palace, located in the ancient city of Nimrud. An image set of the stone reliefs was taken, and used as input for structure-from-motion (SfM) software to create 3D models. These models were then placed into a 3D scene. The scene was designed to provide users with a variety of ways to interact with the space and the stone reliefs. For instance, users can use different lighting models such as sunlight or torchlight to visualize what the reliefs may have looked like in lighting conditions of that time period. In addition, users have the option of enabling an “exaggerated shading” effect that brings out more details in the relief than in possible from viewing them in person. 

 

In the end, it was determined that it was possible to create models with very high geometric resolution using the chosen workflow. Furthermore, as 3D scanning techniques and reconstruction techniques used in this project were deliberately chosen to be both non-contact and easily accessible, this workflow could be used in the future to create many more high quality models of these stone reliefs, and to hopefully reconstruct enough reliefs to fully recreate the experience of the Northwest Palace.

You can see the project at: https://nl384.github.io/nimrud_relief/nimrud.html

sunlight.jpeg
Source: https://nl384.github.io/nimrud_relief/nimrud.html

The 2019 Summer Grant Recipient - Lukas Burger

Lukas, a junior at Yale, is our recipient of the 2019 Summer Grant. Every single person on the Selection Committee gave Lukas a score of 5 - the highest score possible. Below are some answers from Lukas’s application.

What problem do you want to address? Are you working with another Yale person (professor, student, researcher)? 

During my time at Yale, I’ve put most of my energy into the Yale Undergraduate Prison Project. Now as the director of mentoring, I work with incarcerated people from the ages of 18 - 22 every Friday. We do everything from discuss personal issues in their lives to writing down life plans and thinking about jobs. One of the main problems for these people is finding a job that will accept them. I’ve heard countless stories of rejection after rejection. For people who spent their entire time in prison doing productive things like taking classes, working, and attending mentoring sessions this feels very wrong. The people I meet with are all incredibly hard-working, passionate, and intelligent—they really want to do something meaningful in the world, just like any Yale student. But they have grown up in an environment where there are no opportunities and so have ended up in an unfortunate position. As a Yale student, I think it is my duty to help these people, boosting them up as I would anyone else. The main area in which these people struggle, is during the time following their release. They have nothing, many cannot return home, and they do not know if any jobs will accept them. It is during this period that my club and other reentry organizations attempt to help them, but our aid is limited in reach and resources. It is my hope that every newly released person could skip that time of uncertainty and move into a period of personal growth so that one day they can achieve the goals they have spent countless hours dreaming of. I’ve spent every week for the last three years talking with these people about their goals and I’m committed to making them happen. In addition to my time doing mentoring, I’m also the President of Yale Code4Good an organization on campus which works with non-profits to build technical projects. I know first hand what it’s like to build high impact computer science projects. In this project I will be fully supported by my fellow students in the Yale Undergraduate Prison Project, I would rely on our faculty advisor from Dwight Hall, Zelda Roland, and I would work John Mele, our liaison at the Family Reentry non-profit organization. This organization works with incarcerated and at-risk populations across Connecticut and would be a very valuable resource in the area.

What is your proposed solution?

Yale students have the office of career strategy and a job application portal, called Simplicity. I propose a similar solution but for incarcerated people. This web app would contain various resources like resume builders, interview guides, as well as general advice for recently incarcerated job-searchers. The most important section of the website will be the job search section. This would feature jobs which have explicitly stated that they accept formerly incarcerated individuals. After my conversations with many inmates, they say finding a job which will accept them is extremely difficult given their record. My solution would streamline this process, creating a direct route for people to get on the right track. John Mele of Family Reentry has already provided me with a long list of jobs he normally directs his students to, so I have a good place to start from. Each incarcerated person I’ve spoken to has a cellphone or computer, making them ideal users of this platform. They have also expressed to me that they would definitely use a website like this if it existed. The last component of this project would be a texting service, which continually reminds users of new jobs in the area, interviews they have coming up, and when they should follow up with a company. In this way, users will be prompted to keep going along the process, get a job, and set themselves up for success with long term life plans, rather than sinking back into crime. I would first target this application to users in Connecticut, most importantly the cities of New Haven, Bridgeport, and Hartford, as these cities have a very high incarceration rate. I would use the job lists given to me from Family Reentry as a starting point and then expound upon it with my own leads. I would also reach out to members of the department of corrections in CT who have expressed interest in advertising such a platform. This project would distribute the work of various reentry organizations and deliver help to a wide audience. I truly believe this could make change on a wide scale and help these people reach their goals.

What do you hope to learn and accomplish by the end of the summer?

By the end of the summer I will have constructed a service from the ground up which can truly deliver impact and a positive results for the user. I’m excited by this project because it encompasses the whole range of problem solving, including user research, designing, building, and operations. I’m confident I can accomplish this over the summer. I hope to have learned how to build and deploy a fully scalable application and I plan to onboard the first 50+ users this summer. This is achievable given my close ties to CT Family Reentry and the CT department of corrections, who John Mele has introduced me to. I’ve spent my life building various successful tech projects—from a GPS dog tracker in high school, to last year building a crypto currency app and a company to support it, but I’ve never built something which I truly think could make difference in someones else’s life. By the end of the summer, I plan on accomplishing that.

Adriel Sumathipala - Summer Grant 2018 Recipient

Accumulus Bio is a web application that runs computational biology models in the cloud with the click of a button. This year, biotech companies will spend $158 billion on drug development. Yet, drug discovery is still conducted the same way it has been for decades—through repetitive a trial and error process for thousands of potential drug compounds. With the power of cloud computing, Accumulus Bio aims to cut the time spent on early-stage drug discovery by at least 50%.

Accumulus Bio works by running published, well-validated computational biology models to quickly identify the key effects a drug has on a target organ system. Instead of running thousands of guess-and-check experiments that do not return meaningful results, these computational models can quickly identify the few key experiments researchers should run to decide whether to continue with a drug candidate. As a cloud-based platform, Accumulus Bio removes the complexity of running biological models on distributed computing systems. Instead, researchers simply upload their data and with the click of a button run complex biological models on Accumulus Bio’s cloud architecture. Combining the latest advances in computational biology with cloud computing, Accumulus Bio can accelerate drug development and help discover the next-generation of therapeutics

Screen Shot 2019-01-30 at 9.44.31 AM.png
Source: http://www.onlinemusclesim.org/

Flynn Chen - Summer Grant 2018 Recipient

Flynn worked on a Bioinformatics project in the Gerstein Lab at Yale’s Molecular Biophysics and Biochemistry Department.

Flynn says:

I performed analysis on the ENTex data set, and the results are included as a part of the ENCODE consortium. I constructed a pipeline that could calculate the alternative splicing events and performed hierarchical clustering based on these alternative splicing events. The conclusion is that there are more cell-type differences than individual differences in terms of alternative splicing. I also found specific histones that were strongly correlated with different alternative splicing events using results from the pipeline.

I also trained a recurrent neural network (RNN) with E. Coli and parts of the human genome, and the RNN was able to then output snippets of sequences that resemble genes, which were able to be mapped back to the genome of the respective organism. This results of this project could be seen in this link: https://github.com/flynn-chen/Torch-RNN-Trained

flynn.png
Source: https://github.com/flynn-chen/Torch-RNN-Trained

Recipients of the 2018 Summer Grant Selected!

The alums were hard at work reading through all the applications current students submitted for the 2018 Summer Grant.  We are pleased by the diversity of proposed projects.  Our only regret is not being able to offer more students the summer grant.

Originally, we were only able to offer the grant to two students, but through an anonymous one-time donation, we are now able to sponsor six students with full and partial grants of $3,500 to $5,000 each!

Without further ado, here are the recipients of the 2018 Summer Grant from the Andy Keidel Fund at Yale!

 

Flynn Chen (Class of 2020)

I will be working in the Gerstein lab with my mentor Jing at the Department of Molecular Biophysics and Biochemistry on developing a pipeline that would convert raw, unmapped RNA sequences from cancer patients to mapped RNA sequences for analysis as a part of the ENTex consortium. The ENTex consortium is aimed to compare genes expressed in humans, specifically RNA, to previously mapped genome from the ENCODE project. Our contribution to the consortium will be an application of the research in cancer biology. Using high-throughput sequencing data from clinical tumor biopsies, we can compare the difference in the patient's genetic material (DNA and RNA) with already established databases. The pipeline to be developed mostly consist of tools optimized to map reads to reference, sort and mark duplicates, split RNA-specific-overhanging exon regions in the reads, base quality score recalibration, and variant calling. Further differential analysis would then be conducted on the variant-called RNA reads. This genome-wide epigenetic study will shine a new light on post-transcription RNA-editing and its role in oncogenesis. 

 

Roy Elzur (Class of 2019)

The goal of my project this summer is to reduce food waste in the residential college dining halls. To accomplish this goal, I will use data obtained from Yale Hospitality and individual dining halls from past years in order to predict how many students will be eating in a given dining hall on a given day for a given meal. I will train a machine learning model or a neural net on this data, in addition to other potentially useful predictors such as class enrollment and location (especially useful for lunch where people tend to eat at the most conveniently located dining hall given their class schedule), weather (students may prefer to stay close to their college in poor weather), and annual campus events. This project would use R or Python machine learning libraries such as Tensorflow and would ideally also have a simple graphic interface so that dining hall staff may conveniently use it.

An additional component of the project would incorporate IoT scales for determining what percentage of diners eat from each menu item and what the average portion size is. The scales would sit under the food serving containers and would be able to calculate the size of the average food serving by measuring the weight difference each time food is removed from the container. Ideally, the data collected by these scales would be stored in a database and analyzed.

Together, these two components would help dining halls more accurately predict how many students will be eating in a dining hall for a given meal, how popular each menu item is, and what the average serving size for each item is, ultimately giving staff a better idea of how much food needs to be ordered and prepared.

 

Nadia Irwanto (Class of 2020)

I will be working with a team of epidemiologists, decision scientists, mathematical modelers and health economists at the Yale School of Public Health to develop agent-based simulation models of tuberculosis (TB) and HIV that can capture the complexities and dynamic nature of disease epidemics. These models can be used to evaluate the health and financial outcomes of health policies and control interventions, such as contact tracing, before they are implemented, and will hopefully be scalable across any disease in the future. A critical step is to calibrate the models against various demographic and epidemiological data. This is a computationally intensive process and my focus will be to 1) parallelize the existing calibration methods, and 2) design “smarter” calibration algorithms that search the space of parameters and identify the best value in a more efficient way.

 

Nathan Lin (Class of 2019)

The ancient city of Nimrud, near present-day Mosul, Iraq, was the capital of the Assyrian Empire from around 883 BC - 721 BC.  The Northwest Palace was the royal residence and center of government during this period, and many rooms were decorated with intricate stone reliefs.  As excavations have been carried out intermittently since 1845, these reliefs are now housed in many museums around the world, including in the Yale University Art Gallery. Unfortunately, many of these reliefs remain at the Northwest Palace where they are believed to have been destroyed by the Islamic State.

My project is centered on 3D scanning and rendering of the stone reliefs held at the Yale University Art Gallery. One of the main focuses of my project is to be able to create a virtual reality (VR) viewing experience for users at home. I hope to be able to present to users a very realistic portrayal of how the stone tablets may have looked like under the lighting/coloring conditions of the Northwest Palace. In addition, I hope to be able to present to the viewer a wide variety of different coloring and lighting hypotheses that can be quickly and intuitively cycled between using VR gestures.

 

Gabriel Saruhashi (Class of 2020)

Project ReConnect aims to bridge a critical knowledge gap between refugees and their host communities by providing a virtual two-way learning platform that connects a local with a newcomer prior to their arrival in the country of relocation. ReConnect addresses the lack of pre-departure/arrival education and information of the two communities whilst also enabling refugees to create and develop genuine networks within countries of relocation. This initiative helps reduce social stigmas and negative stereotypes associated with refugees and promotes better understanding of global forced migration challenges.

By the end of the summer, I will have finished the beta version of the website and mobile app. My key metric of success will be to promote at least 100 ReConnections over the summer. This means that I hope to welcome more 100 host friends and 100 refugees to the ReConnect community, who will be actively using both the browser and mobile products. With the testimonials and videos we collect in Paris, I also hope to create a blog section on the website to share ReConnect stories and raise awareness on the issues of forced migration, social integration and diversity.

 

Adriel Sumathipala (Class of 2020)

Acumulus is a web application that runs computational biology models in the cloud with a click of a button. This year, biotech companies will spend $158 billion on drug development. Yet, drug discovery is still conducted the same way it has been for decades—through repetitive a trial and error process for thousands of potential drug compounds. With the power of cloud computing, Acumulus aims to cut the time spent on early-stage drug discovery by at least 50%.

Acumulus works by running published, well-validated computational biology models to quickly identify the key effects a drug has on a target organ system. Instead of running thousands of guess-and-check experiments that do not return meaningful results, these computational models can quickly identify the few key experiments researchers should run to decide whether to continue with a drug candidate. As a cloud-based platform, Acumulus removes the complexity of running biological models on distributed computing systems. Instead, researchers simply upload their data and with a click of a button run complex biological models on Acumulus’ cloud architecture. Combining the latest advances in computational biology with cloud computing, Acumulus can accelerate drug development and help discover the next-generation of therapeutics.

Learn to Program on the Go!

Ever been curious about programming but don’t know where to start?  The Andy Keidel Fund is excited to announce our partnership with Mimo, an interactive mobile app that teaches computer programming in a fun way. Anyone with a Yale email can sign up for a free one-year premium Mimo service by following this link: https://getmimo.com/yale

Give it a try, and let us know what we can improve by sending us a message.

Fully funded

In December 2016, the family and friends of Andy Keidel officially established the Andy Keidel Fund through an endowment at Yale University. More information about student projects and ventures will be available in the coming months. 

Going live

Our website is newly launched. Please check back frequently for news and updates.