Murthy Devarakonda

Samuel C Johnson Research Bldg, Mayo Clinic
13212 E. Shea Blvd
Research Professor, MY


My educational background is in computer science. I received Ph.D. in computer science from University of Illinois at Urbana-Champaign, Illinois. All of my professional experience until joining ASU was as a Research Scientist and as a team manager at IBM Research, Yorktown Heights, NY.

I was a member of the well known IBM Watson team (who created IBM Watson that defeated two legendary champions in a televised Jeopardy! quiz show in 2011). I joined the team after the televised event in 2011 and through 2017, I was the lead and the team manager for an initiative to apply IBM Watson methodology to longitudinal patient records, especially concentrating on the textual narratives of the patient records. We worked with patient records and physicians from multi-specialty hospitals, developing natural language processing and machine learning methods for automatically extracting insights such as the problem list from the text narratives of a patient record.

Prior to the IBM Watson tenure, my research was mainly in distributed computing systems where I applied data science (before it was called by that name) to identify insights from measurement data and to use the insights as foundational elements to develop novel algorithms and distributed systems.

I am a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), and a Distinguished Member of the Association for Computing Machinery (ACM). I received an Outstanding Technical Achievement Award, a Master Inventor award, and several Research Division Awards at IBM. I published over 60 refereed papers in conferences and journals. I received 42 patents for the work at IBM Research.

Google Scholar

Research Interests

  • Natural Language Processing
  • Data Science
  • Machine Learning
  • Cloud Computing
  • Distributed Computing and Data


Bioinformatics Publications

  1. Bharath Dandala, Venkata Joopudi, Murthy Devarakonda, “Adverse Drug Events Detection in Clinical Notes by Jointly Modeling Entities and Relations using Neural Networks”, Drug Safety, 2019 (to appear).
  2. Venkata Joopudi, Bharath Dandala, Murthy Devarakonda, “A Convolutional Route to Abbreviation Disambiguation in Clinical Text”, Journal of Biomedical Informatics, October 2018.
  3. Neil Mehta, Murthy Devarakonda, “Machine Learning, Natural Language Processing and Electronic Health Records: The Next Step in the Artificial Intelligence Journey”, Journal of Allergy and Clinical Immunology (the official publication of the American Academy of Allergy, Asthma, and Immunology), 2018 (to appear)
  4. Murthy Devarakonda, Neil Mehta, Ching-Huei Tsou, Jennifer Liang, Amy Nowacki, and John Eric Jelovsek, “Automated problem list generation and physicians perspective from a pilot study” Intl Journal of Medical Informatics (Elsevier), Vol 105, pp 121-129, Sept 2017.
  5. Murthy Devarakonda, Neil Mehta, MD: Cognitive Computing for Electronic Medical Records, in Healthcare Information Management Systems, 4th Edition, Springer International, October 2015.
  6. Bharath Dandala, Diwakar Mahajan, Murthy Devarakonda, “IBM Research System at TAC 2017: Adverse Drug Reactions Extraction from Drug Labels”, (invited paper), Text Analysis Conference (TAC) 2017 Workshop at NIST, Gaithersburg, MD, USA, Nov 2017.
  7. John Prager, Jennifer J Liang, Murthy Devarakonda, "Watson SemanticFind: Locating what you want in an EMR, not just what you ask for", AMIA Joint Summits, Mar 2017
  8. Jennifer J Liang, Ching-Huei Tsou, Murthy Devarakonda, "Ground Truth Creation for Complex Clinical NLP Tasks – An Iterative Vetting Approach and Lessons Learned" AMIA Joint Summits, Mar 2017
  9. Homa Alemzadeh, Murthy Devarakonda, "An NLP-based Cognitive System for Disease Status Identification in Electronic Health Records,” IEEE EMBS – BHI Conference, Feb 2017
  10. Bharath Dandala, Murthy Devarakonda, Mihaela Bornea, Christopher Nielson: Scoring Disease-Medication Associations using Advanced NLP, Ma-chine Learning, and Multiple Content Sources, 5th Workshop on Building and Evaluating Resources for Biomedical Text Mining (BioTxtM) held with COLING 2016, Nov 2016, Japan
  11. Ching-Huei Tsou, Murthy Devarakonda, Jennifer J. Liang: Toward Generating Domain-specific / Personalized Problem Lists from Electronic Medical Records, AAAI Fall Symposia, Nov 2015.
  12. Murthy Devarakonda, Ching-Huei Tsou: Automated Problem List Generation from Electronic Medical Records in IBM Watson. AAAI Winter Conference – Innovative Applications of Artificial Intelligence Track, Austin, Texas, Jan 2015
  13. Puripant Ruchikachorn, Jennifer J. Liang, Murthy Devarakonda and Klaus Mueller: Watson-Aided Non-Linear Problem-Oriented Clinical Visit Preparation on Tablet Computer, IEEE EHR VIS 2014 Workshop, Paris, Nov 2014
  14. Murthy Devarakonda, Dongyang Zhang, Ching-Huei Tsou, Mihaela Bornea: Problem-Oriented Patient Record Summary: An Early Report on a Watson Application, IEEE HealthCom, Natal, Brazil, Oct, 2014.

47 additional publications on data science for distributed computing and storage (see

Research Activity

My current research interests are in exploring natural language processing (NLP) and machine learning for medical and clinical applications. A large portion of a patient record is clinical narratives, known as clinical notes, written by care providers such as physicians and nurses. The clinical notes include information about the patient and their treatment beyond what is included in the structured and coded data, and so it's a gold mine for knowledge and patient care insights. Novel NLP methods can unlock the value from the patient records by improving efficiency and effectiveness of care providers. treatment processes, adverse drug reactions, and drug repurposing to name a few.

The number of publications and clinical trial data is increasing at an exponential rate, and novel methods are needed to bring this information to physicians, researchers, policy-makers, and patientsin in a usable form. Social media including blogs and tweets are another great source of insights into patient care and outcomes. Internet of Things (IoT) devices collect large quantities of data about us, often unitrusively and without participant awareness,  which can provide new insights into individual and group health, when combined with patient health records. My research explores these sources of data for deep insights useful in health and wellness care.

Exciting new advances in neural networks methds for NLP accelerate the process of extracting insights from vast medical and clinical text sources. A growing number of tools available to use the new methods offer a low-barrier to entry in applying the novel methods. Increasing availability of low cost cloud computing and GPU technology provides the necessary computing platforms for the adavced methods. My research includes exploring these foundational aspects for improved healthcare.

Recent and Ongoing Research Projects:

(1) NLP-based Automated Cohort Selection from Clinical Notes

Team tied to 1st place at National NLP Clinical Challenges shared-task for Cohort Selection

(2) Improved neural network methods for adverse drug events extraction from clinical notes

(3) Neural networks based Learn2rank to find patient-specific treatments from scientific literature

(4) Identifying complex events from inexpensive sensor data from living spaces


Spring 2019
Course NumberCourse Title
BMI 484Internship
BMI 493Honors Thesis
BMI 505Foundations of BMI Methods II
BMI 560Teachng Biomedical Informatics
BMI 593Applied Project
BMI 790Reading and Conference
BMI 792Research
BMI 799Dissertation
Fall 2018
Course NumberCourse Title
BMI 492Honors Directed Study
BMI 560Teachng Biomedical Informatics
BMI 593Applied Project
BMI 598Special Topics
BMI 799Dissertation
Spring 2018
Course NumberCourse Title
BMI 484Internship
BMI 493Honors Thesis
BMI 499Individualized Instruction
BMI 505Foundations of BMI Methods II
BMI 560Teachng Biomedical Informatics
BMI 593Applied Project
BMI 790Reading and Conference
BMI 792Research
BMI 799Dissertation
Summer 2017
Course NumberCourse Title
BMI 593Applied Project


Spring 2019: BMI 505 Methods II (Intro to NLP, Database Modeling & Querying, Intro to KRR)

Fall 2018: BMI 598 - Special Topics: NLP Methods for Biomedical Text Mining

Spring 2018: BMI 505 Methods II (Intro to NLP, Database Modeling & Querying, Intro to KRR)