Ara Vartanian
PhD Candidate
Department of Computer Sciences
email: first name and first four letters of my last name no spaces nor punctuation nor nuthin at cs dot wisc dot edu
Is the above obfuscation a bit much? Okay, it's aravart. Show me what your crawler's got!
About Me
I am a PhD candidate in the Department of Computer Sciences, University
of Wisconsin-Madison, working with advisor Xiaojin (Jerry) Zhu. My research focuses on interactive machine learning (iML), which is machine learning with a human-in-the-loop. In traditional machine learning, a human provides input to a learning system only through labeling examples.
iML explores the improved learning outcomes possible when there is a richer interface between human and computational agent. Sometimes, for example, we don't have a lot of data at our disposal, but a human can efficiently guide a learner through a small number of specially constructed examples. Conversely, sometimes a computational system can aid a human in quickly synthesizing a large number of training examples for then training a computational learner.
From a theoretical standpoint, my research frames interactive machine learning within the context of machine teaching and active learning, touching on ideas and methods in reinforcement learning and optimal control theory. Since adversarial attacks are the flip side of machine teaching, I’ve also done work in adversarial learning. Along the way, I’ve worked on applications in NLP, program synthesis, and topological data analysis.
Publications
- Preference-Based Batch and Sequential Teaching: Towards a Unified View of Models. Farnam Mansouri, Yuxin Chen, Ara Vartanian, Xiaojin Zhu, Adish Singla. NeurIPS, 2019.
- Attacking Data Transforming Learners at Training Time. Scott Alfeld, Ara Vartanian, Lucas Newman-Johnson, Benjamin Rubinstein. AAAI, 2019.
- Training Set Camouflage. Ayon Sen, Scott Alfeld, Xuezhou Zhang, Ara Vartanian, Yuzhe Ma, and Xiaojin Zhu. GameSec, 2018.
- Program Synthesis with Visual Specification. Evan Hernandez, Ara Vartanian, and Xiaojin Zhu. ArXiv 1806.00938.
- Model-Assisted Machine-Code Synthesis. Venkatesh Srinivasan, Ara Vartanian, and Thomas Reps. OOPSLA, 2017.
- Active Learning with Oracle Epiphany. Tzu-Kuo Huang, Lihong Li, Ara Vartanian, Saleema Amershi, and Xiaojin Zhu. NeurIPS, 2016.
- Stochastic Multiresolution Persistent Homology Kernel. Xiaojin Zhu, Ara Vartanian, Manish Bansal, Duy Nguyen, and Luke Brandl. IJCAI, 2016.