Summer Undergraduate Research Fellowships

SURF—the Summer Undergraduate Research Fellowships program—is the centerpiece of Smith’s Summer Research Fellowship program. Hands-on research collaboration between faculty and students is a cornerstone of science education at Smith, so the summer program has always been very important to us.

Who Is Hiring?

See the projects for summer 2019 to get an idea about the exciting research work happening at Smith.

Alicia M. Grubb

This summer I’m looking for students interested in software development and/or empirical research (i.e., Software Engineering).

• In past summers, we developed BloomingLeaf, a web-based tool for building and analyzing goal models. This summer we will focus on improving the usability of our analysis results by creating better visualization and smart algorithms to help stakeholders understand trends over time. We will also implement new model management features helping multiple stakeholders compare their models and results. We seek the help of enthusiastic students with strong programming skills and an exposure to web development (javaScript), who have excellent communication skills (written/verbal English) and an ability to work independently and collaboratively. Our online codebase will help you create a portfolio for future applications to employers and graduate schools.

• Other projects include: using stakeholder preferences to reduce the solution space of possible evolutions; comparing expressive power and usability of goal modeling languages; visualizing trends in evolutionary reasoning; and exploring the utility of goal modeling activities. Students with an interest in research and/or an interest in qualitative data analysis (e.g., open coding, ethnography) should state this on their application.

R. Jordan Crouser

• The Automated Scribal Identification Project uses ancient manuscripts written in the Aramaic dialect of Syriac as a case study for exploring how recent advances in the digital analysis of handwriting can help scholars better ascertain a manuscript's provenance, identify manuscripts written by the same scribe, and trace out the chronological development of ancient scripts. Although the initial project goal is to substantially advance our understanding of Syriac Christianity, its greatest effect would be in providing a platform and a model for similar ventures in other languages such as Arabic, Greek, Hebrew, Latin, and Sanskrit.

• Individual Differences Project: Researchers have conducted many studies that observe a data analyst performing a particular task, and have used those observations to define an analytic workflow. Individual analysts have distinctive workflows: they may combine canonical actions in unique patterns, focus on particular areas over others, and so forth. But, how do we predict which workflows an analyst is likely to favor in advance? In this project, we will design and conduct a series of experiments aimed at understanding the invariant features of the user that can be used to identify which features of an analytic tool we can modulate to better support individual analyst. By mapping features of theuseronto features of the tool, we aim to provide a better, more streamlined experience for the user, thereby amplifying human analytic capabilities in areas where purely computational analysis fails.

Katherine M. Kinnaird

This summer, I am looking for students interested in building a new Python package and/or interested in building an interactive visualization platform for a representation of musical scores. Additionally, I may have funding for a student interested at the intersection of data science education research and text mining.

• Building a new Python package that reimagines existing MATLAB code for building aligned hierarchies, a representation for musical scores. This builds on work published at ISMIR 2016, but requires no background knowledge in music information retrieval (MIR). Students interested in this project should have experience coding in Python (ie. CS 111), and they should be excited by the prospect of building code that will be publicly available.

• The Interactive Aligned Hierarchies also builds on the aligned hierarchies, but links the static output to the score. This will allow for interactive exploration of both the score and the representation simultaneously. Students interested in this project should have experience designing large-scale projects and are excited about connecting visualizations with sound files.

• (Possible funding) As part of a TRIPODS+X grant, there is summer work to explore investigating students’ data science misconceptions before and after their first formal course in data science. This project will rely on aspects of text mining in addition to data management.

John Foley

Pending SURF funding for lecturers. I have three projects I will be working on this summer with opportunities for undergraduate participation.

• Poetry Project. I have the largest digital collection of poetry in the world, but there is little organizational information about it. I am looking for someone interested in identifying genre or form in these documents of identified poetry: which are sonnets, which are parts of plays written in verse, etc. Other directions of curation of this dataset are possible.

• A Programming Language for Information Retrieval (Search). The primary users of search systems in the future will be autonomous agents, like digital voice assistants or other intelligent systems that help users with complex tasks. Students interested in programming languages, compilers or information retrieval are encouraged to discuss project ideas with me.

• Explainable Reinforcement Learning. Reinforcement Learning agents are being developed for autonomous vehicles and other safety-critical applications. Unfortunately, modern models for this type of learning struggle to explain their predictions even if their "accuracy" is much higher than more traditional models. Students interested in deep reinforcement learning (pytorch, tensorflow, etc.) could work with us to implement recent models and students interested in automated testing could work with us to implement decision explanation scripts -- answering "what if?" questions about model behavior.

Sahar Al Seesi

Pending SURF funding for lecturers.

• GeNeo is a suite of tools used to predict short protein sequences (neo-epitopes) that can be used as cancer vaccines in personalized cancer therapy. Interested students could help in developing the next version of this suite of tools. This work will give you the opportunity to gain some biological background and be part of exciting work that is currently used in an ovarian cancer clinical trial! Most of the work will be done in shell scripting (which you can learn during the summer) and Python or Java.

• Allele Specific Gene Expression Estimation is an interesting computational biology problem that answers the question of whether the paternal, the maternal, or both copies of an inherited gene are expressed (active) in your body and at what level they are expressed. This question is asked by many biologists, but the computational methods to answer it are still not good enough to accurately give an answer. In this research project, we will build software to address this problem by combining ideas in existing research with some new research ideas.

Nick Howe

Handwriting recognition and historical manuscripts. Read more about this research of Nick Howe’s website.

The summer of 2019 marked the 51st year that Smith has had a formal student summer research program! We invite you to join the students who will receive stipends to support their independent research during the summer.

APPLY