Educational Activities
- Tutorial on Machine Learning. In this comprehensive tutorial we provide introductions to several important machine learning methods and several case studies of their application in biomedicine. The tutorial contains among other topics a very accessible and concise introduction to Support Vector Machines, Markov Blanket and Causal Discovery Algorithms. A broad range of resources (bibliography, links and software) is also provided.
- Classes. The DSL faculty offers, or participates in, many classes offered at Vanderbilt.
- Opportunities for formal studies: Degree program. The Vanderbilt Degree Program in Biomedical Informatics is part of the Faculty of Arts and Sciences offerings and leads qualified candidates to the M.S. and Ph.D. degrees in Biomedical Informatics. Qualified students who are US citizens or permanent residents may be eligible for NIH/NLM support through the Department of Biomedical Informatics NIH Training grant.
- Opportunities for mentored research studies: Non-Degree program. The Vanderbilt Department of Biomedical Informatics non-Degree Program is offered to qualified post-doctoral candidates and leads to a structured non-degree Fellowship in Biomedical Informatics. Qualified trainees who are US citizens or permanent residents may be eligible for NIH/NLM support through the Department of Biomedical Informatics NIH Training grant.
- Qualifications of prospective graduate students and post-doctoral fellows. Here we describe the prior training and experience sought in students interested in working with DSL faculty mentors.
- A list of academic & training programs in Biomedical Informatics in the USA.
- BMIF 315. Methodological Foundations of Biomedical Informatics. In this course, students will develop foundational concepts of computation and analytical thinking that are instrumental in solving challenging problems in biomedical informatics. The course will use lectures and projects directed by co-instructors and a cohort of guest lecturers. Instructors: D. Giusea and C.F. Aliferis. [course homepage]
- BMIF 330. Biomedical Artificial Intelligence I. Part 1: Decision-Support Systems. Fundamentals of AI programming. Search algorithms. Overview of Propositional and First Order Logic (FOL). Reasoning in FOL (Generalized Modus Ponens, Forward and Backward Chaining, Resolution). Formal reasoning systems. Applications of Description logics in biomedical informatics. Early Bayesian and ad-hoc systems for medical diagnosis and decision making. Clinical Decision Support Systems Knowledge Bases. Medical knowledge/data acquisition and maintenance for Decision Support. Bayesian Networks.
Part 2: Machine Learning. Introduction to ML programming techniques. Data cleaning and preparation Machine Learning Inductive framework. Mathematical foundations. Algorithm families: Decision Tree Induction, Genetic Algorithms, Neural Networks, Clustering, K-Nearest Neighbors, Support Vector Machines, Feature Selection, Causal Discovery methods with Bayesian Networks. Instructors: C.F. Aliferis and S. Mani
[course homepage]
- BMIF 330A. Medical Artificial Intelligence I Laboratory. Applications and in-depth study of algorithms and topics introduced in Medical Artificial Intelligence I.
Fundamentals of AI programming with LISP. Implementation of Search algorithms, unification, Forward and Backward Chaining. Representation in Semantic Networks and Description logics. Implementation of the Internist I algorithm. Introduction to ML programming techniques with MATLAB. Implementation and testing of various machine learning algorithms.
Instructors: C.F. Aliferis and S. Mani. [course homepage]
- BMIF 360. Graduate Seminar on Biomedical Informatics Algorithms. Graduate-level topics in intermediate or advanced algorithms, data structures and knowledge representations for biomedical informatics that are not covered in the M.S./Ph.D. core courses. Examples include: Bayesian learning of Causal Probabilistic Networks, applications of the EM family of algorithms, advanced clustering methods (Machine Learning); Constraint-satisfaction search/programming for optimization, Markov Decision Processes, Hidden Markov Models (Artificial Intelligence); Controlled Vocabularies, statistics for content-bearing terms, efficient storage and retrieval of lexical terms, citation analysis and the PageRank algorithm, collaborative filtering, automatic text categorization (Information Retrieval); exact and applied sequence comparison, algorithms for computational proteomics (computational biology); mathematical techniques from probability, linear algebra, and optimization theory. Note: covered topics will be highly dependent on faculty and students interests and will change from year to year to reflect research advances and interests. Students must first obtain instructor permission to enter the class.
- CANB 342. Cancer Biology. Advanced concepts in Cancer Biology will be reviewed using a combination of lectures, and discussion sessions based on current literature. Topics range from the molecular biology of cancer (oncogenes and tumor suppressors) to issues of drug design and clinical trials. Prerequisite: IGP core course or consent of instructor. [Fall 2005 materials], [Fall 2006 materials], [Fall 2007 materials]
- MSCI 524-5026. Proteomics. Introduction to Clinical Proteomics—rationale, description of prior studies and future needs. Selection of candidate biomarkers—application of MALDI Mass Spec to biological samples. Selection of candidate biomarkers—other approaches. Assay development – multiplex assays, high throughput micro assays, industry collaboration. Sample datasets—how clinical proteomics can be incorporated into past and future clinical trial datasets and IRB issues. Biostatistical Analytical approaches for proteomics. Novel Bioinformatic approaches to proteomic data analysis. Clinical Proteomics in action—application of the process to clinical acute lung injury.
The Department of Biomedical Informatics (DBMI) offers MS and PhD degree programs. Combined MS/MD and MD/PhD programs are also offered. These are programs of intensive study that balance coursework and cutting-edge research. DBMI has an NIH training grant that (on a competitive basis) funds qualified students who are US citizens or permanent residents (Green Card holders). Prospective students who may be interested to work in the DSL research areas should apply to one of the above programs directly at the DBMI website.
In addition to the MS and PhD degree programs, DBMI also offers NIH-funded post-doctoral research as well as shorter internships. Post-doctoral
research typically lasts from two to three years. Short internships are typically one semester long.
Prospective post-doctoral fellows and interns who may be interested to work in the DSL research areas
should apply directly to DBMI.
Many graduate students and
fellows elect to work in areas other than those pursued within DSL. Typically a student is encouraged to
have some areas of strong interest but also be flexible enough to decide on a career path after he or she
has taken courses during the first two semesters in the program and been exposed to the spectrum of
possibilities that the field has to offer. It is not uncommon for students to enter with a strong clinical systems interest
and end up doing bioinformatics work and vice versa.
Especially with respect to students wishing to work with DSL faculty the following are considered strong qualifications:
- strong familiarity with fundamental data structures and algorithms,
- non-trivial programming experience in one or more general-purpose computer languages,
- strong quantitative skills and mathematical aptitude,
- familiarity with fundamental concepts and methods of biomedical research and practice,ideally combined with some prior research or industrial experience that substantiates academic potential in the field.
- working knowledge of networking, statistics, and databases,
- strong academic record, recommendations, and standardized scholastic aptitude tests, and
- excellent command of written and spoken English (for foreign prospective students).
Copyright © 2002-2007, Discovery Systems Laboratory, Department of Biomedical Informatics, Vanderbilt University, Nashville, TN, USA