Research :: Major academic projects and thesis work.


Kidney Exchange.

2010—Present

The preferred treatment for kidney failure is transplantation. However, the demand for donor kidneys is far greater than supply. In the US, roughly 35,000 people are added to a waiting list of approximately 100,000 each year, while only 16,000 leave due to receiving a kidney. Demand is increasing worldwide.

Kidney exchange is a recent innovation that allows patients who suffer from terminal kidney failure, and have been lucky enough to find a willing but incompatible kidney donor, to swap donors. The algorithms and code we develop at CMU currently run the US nationwide kidney exchange program (facilitated through UNOS); this program now includes over 140 transplant centers and matches on a biweekly basis.

There are many open problems in kidney exchange—in operations research, economics, policy, and medical data mining. Feel free to get in touch if any of these strike your fancy!

Representative Publications

  • Dickerson, J.P., Kazachkov, A.M., Procaccia, A.D. and Sandholm, T. 2017. Small Representations of Big Kidney Exchange Graphs. Conference on Artificial Intelligence (AAAI) (2017). [link]
  • Dickerson, J.P., Manlove, D., Plaut, B., Sandholm, T. and Trimble, J. 2016. Position-Indexed Formulations for Kidney Exchange. Conference on Economics and Computation (EC) (2016). [link]
  • Dickerson, J.P. and Sandholm, T. 2015. FutureMatch: Combining Human Value Judgments and Machine Learning to Match in Dynamic Environments. Conference on Artificial Intelligence (AAAI) (2015). [link]
  • Dickerson, J.P., Procaccia, A.D. and Sandholm, T. 2014. Price of Fairness in Kidney Exchange. International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS) (2014). [link]
  • Dickerson, J.P. and Sandholm, T. 2014. Multi-Organ Exchange: The Whole is Greater than the Sum of its Parts. Conference on Artificial Intelligence (AAAI) (2014). [link]
  • Dickerson, J.P., Procaccia, A.D. and Sandholm, T. 2013. Failure-Aware Kidney Exchange. Conference on Economics and Computation (EC) (2013). [link]
  • Dickerson, J.P., Procaccia, A.D. and Sandholm, T. 2012. Dynamic Matching via Weighted Myopia with Application to Kidney Exchange. Conference on Artificial Intelligence (AAAI) (2012). [link]
  • Dickerson, J.P., Procaccia, A.D. and Sandholm, T. 2012. Optimizing Kidney Exchange with Transplant Chains: Theory and Reality. International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS) (2012). [link]
Kidney exchange.

Dynamic TV and Cross-Media Advertising.

Dynamic TV and Cross-Media Advertising.

2012—Present

With the increasing popularity of non-linear TV and the introduction of new content delivery mechanisms, monetization in the TV space is becoming more complex. With my advisor Tuomas Sandholm, I am designing a new marketplace for advertising sales that uses optimization to tame this complexity from the human's perspective—while increasing overall market efficiency from a computational economics point of view.

We have pilot programs under way with two of the nation's largest MSOs, and are currently specifying a pilot with a large broadcast TV content provider.

Representative Publications

Coming soon! We're working on the industrial side of things first; check out Optimized Markets for details.


Counter-terrorism.

2008—Present

We have shown that large-scale analysis of terror groups benefits from the use of current computational analysis techniques through a variety of theoretical and real-world case studies, including:

  • an in-depth look at the South Asian terrorist organization Lashkar-e-Taiba, using data collected by social scientists covering 770 features of the group over the last two decades;
  • using geospatial abduction to infer the locations of improvised explosive device (IED) and weapons caches in Baghdad, Iraq, and to recommend search strategies for multiple possible locations;
  • investigating static and dynamic asset protection (e.g., "what is the best way to protect a parade route?") on large graphs against game-theoretic adversaries; and
  • theoretically and empirically analyzing the five-player game including the US, India, Pakistan's government, Pakistan's military, and Lashkar-e-Taiba under the assumption that different policy experts will have different viewpoints regarding various actions taken by each player.

To discuss these projects in greater depth, please get in touch: dickerson@cs.cmu.edu.

Representative Publications

  • Sawant, A., Dickerson, J.P., Hajiaghayi, M.T. and Subrahmanian, V.S. 2015. Automated Generation of Counter-Terrorism Policies using Multi-Expert Input. ACM Transactions on Intelligent Systems and Technology (TIST). (2015). [link]
  • Dickerson, J.P., Sawant, A., Hajiaghayi, M. and Subrahmanian, V.S. 2013. PREVE: A Policy Recommendation Engine based on Vector Equilibria Applied to Reducing LeT's Attacks. International Conference on Advances in Social Networks Analysis and Mining (ASONAM) (2013). [link]
  • Subrahmanian, V.S., Mannes, A., Sliva, A., Shakarian, J. and Dickerson, J.P. 2012. Computational Analysis of Terrorist Groups: Lashkar-e-Taiba. Springer. [link]
  • Shakarian, P., Dickerson, J.P. and Subrahmanian, V.S. 2012. Adversarial Geospatial Abduction Problems. ACM Transactions on Intelligent Systems and Technology (TIST). 3, 2 (2012), 34:1--34:35. [link]
  • Dickerson, J.P., Simari, G.I., Subrahmanian, V.S. and Kraus, S. 2010. A Graph-Theoretic Approach to Protect Static and Moving Targets from Adversaries. International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS) (2010). [link]
  • Subrahmanian, V.S. and Dickerson, J.P. 2009. What Can Virtual Worlds and Games Do for National Security?. Science. 326, 5957 (2009), 1201-1202. [link]
Counter-terrorism image.