Ameet Soni

Ameet Soni

Assistant Professor
Computer Science Department
Swarthmore College

phone: (610) 957-6288
office: 283 North Parrish

I am currently an Assistant Professor of Computer Science at Swarthmore College. I received my Ph.D. in Computer Science in August 2011 from the University of Wisconsin where I was advised by Professor Jude Shavlik. My general research interests are in the areas of machine learning and computational biology and medicine.

Current Semester

2015-16 Schedule
  CS21   Introduction to Computer Science
          Lecture   1:15-2:30pm Tuesday, Thursday 256 Sci Ctr
  Office Hours   10:00am-noon Wednesday, and by appointment 253 Sci Ctr
  Prep time (limited availability)   10pm-noon, Tuesday, Thursday 253 Sci Ctr
  Research Hours   1:00pm-4pm Wednesday, Friday  

Previous Courses:

Research Interests

Selected Publications

[Complete List]

Asterick's (*) indicate supervised students.

Learning relational dependency networks for relation extraction.
, , , .
In Proceedings of the 26th International Conference on Inductive Logic Programming (ILP), . [slides]
[bibtex] [pdf]

A Comparison of weak supervision methods for knowledge base construction.
, , , .
In 5th Workshop on Automated Knowledge Base Construction (AKBC) at NAACL, . [poster]
[bibtex] [pdf]

A comprehensive analysis of classification algorithms for cancer prediction from gene expression.
, .
In Proceedings of the 6th ACM Conference on Bioinformatics, Computational Biology and Health Informatics (ACM-BCB), pp. 525–526, . [poster]
[bibtex] [pdf]

A support program for introductory CS courses that improves student performance and retains students from underrepresented groups.
, , , , , .
In Proceedings of the 45th ACM Technical Symposium on Computer Science Education (SIGCSE), pp. 433–438, .
[bibtex] [pdf]

A graphical model approach to ATLAS-free mining of MRI images.
, , , .
In Proceedings of the 2014 SIAM International Conference on Data Mining (SDM), pp. 974–982, . [poster]
[bibtex] [pdf]

Probabilistic ensembles for improved inference in protein-structure determination.
, .
In Proceedings of the 2nd ACM Conference on Bioinformatics, Computational Biology and Biomedicine (ACM-BCB), pp. 264–273, . Invited for journal publication. [slides] [cached]
[bibtex] [pdf]

Guiding belief propagation using domain knowledge for protein-structure determination.
, , .
In Proceedings of the First ACM International Conference on Bioinformatics and Computational Biology (ACM-BCB), pp. 285–294, . Best Paper Award. [slides] [cached]
[bibtex] [pdf]

Creating protein models from electron-density maps using particle-filtering methods.
, , , , , , .
In Bioinformatics, Oxford Univ Press, vol. 23, no. 21, pp. 2851–2858, . PMCID: PMC2567142
[bibtex] [pdf] [doi]

Improved methods for template-matching in electron-density maps using spherical harmonics.
, , , .
In IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp. 258–265, . Invited for journal publication. [code/data]
[bibtex] [pdf]


Most recently, I have begun researching several different computational problems in the area of MRI braining imaging. Specifically, I seek to apply probabilistic methods to improve the ability to diagnose onset of Alzheimer's disease and other cognitive impairments. Currently, my group is pursuing methods for segmenting anatomical regions of the brain using graph-based methods as well intuit anatomical relationships between patients in the larger ADNI study.

Previously, I worked on ACMI (Automated Crystallographic Map Interpretation). The task of determining protein structures has been a central one to the biological community for several decades. The structure allows biologists to extract information about the underlying biology of a protein, and has implications for various applications such as disease treatment, drug design, and protein design. The most popular method for producing protein structures is by interpreting an electron-density map - a three-dimensional image of a molecule produced through X-ray crystallography. This process, however, remains a resource- intensive and time-consuming task, stunting basic biological research. Thus, the main objective of the project is:

Given the electron-density map (3D image) and a primary sequence of a target protein, produce a three- dimensional, physically-feasible, all-atom model of the target protein's structure.

The result of our group's efforts is ACMI, a probabilistic technique for determining protein structures. Prior to ACMI, techniques failed when trying to interpret low-quality images. With ACMI, crystallographers can now obtain complete and accurate structures from these difficult proteins instead of scrapping the project or dedicating months of effort.

Research Students

Previous Current