Summer 2017–present: Software Engineer, Google Research, New York, NY.
Member of the Structured Data Group in Google Research, working on computational journalism. I use machine learning, natural language processing, and information retrieval to help computers understand news and build products that will make news more informative and engaging.
Worked on shopping ads and recognizing products in images.
Built a system that connects various computer vision tools and
incorporates semantic data to map objects found in images to
visually and semantically most similar product offers.
Worked on Google's click-through rate prediction system.
Designed and implemented algorithms for searching for regions of the feature space
where the click-through rate prediction models exhibit high bias.
Summer 2014: Research Scientist Intern, Amazon, Seattle, WA.
Worked on data center capacity forecasting for Amazon Web Services.
Created a new methodology that is based on machine learning and utilized for long-term capacity forecasting.
Per backtesting evaluations, the new method significantly reduced error compared to the incumbent AWS forecasting process.
Conference Papers
AgreeSum: Agreement-Oriented Multi-Document Summarization (arXiv).
Richard Yuanzhe Pang, Adam D. Lelkes, Vinh Q. Tran, and Cong Yu.
Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021.
Quiz-Style Question Generation for News Stories (arXiv).
Adam D. Lelkes, Vinh Q. Tran, and Cong Yu.
In Proceedings of The Web Conference 2021.
Investigating Rumor News Using Agreement-Aware Search (arXiv).
Jingbo Shang, Tianhang Sun, Jiaming Shen, Xingbang Liu, Anja Gruenheid, Flip Korn, Ádám D. Lelkes, Cong Yu, and Jiawei Han.
In Proceedings of the 2018 ACM on Conference on Information and Knowledge Management
(CIKM 2018).
A Confidence-Based Approach for Balancing Fairness and Accuracy (arXiv).
Benjamin Fish, Ádám D. Lelkes, and Jeremy Kun.
In Proceedings of the 2016 SIAM International Conference on Data Mining
(SDM 2016).
On the Computational Complexity of MapReduce (arXiv).
Benjamin Fish, Jeremy Kun, Ádám D. Lelkes, Lev Reyzin, and György Turán.
In Proceedings of the 29th International Symposium on Distributed Computing
(DISC 2015).
Interactive Clustering of Linear Classes and Cryptographic Lower Bounds (pdf, slides).
Ádám D. Lelkes and Lev Reyzin. In Proceedings of the 26th International Conference on Algorithmic Learning Theory
(ALT 2015).
Journal Papers
Network installation under convex costs (journal).
Alexander Gutfraind, Jeremy Kun, Ádám D. Lelkes, and Lev Reyzin. In Journal of Complex Networks, Volume 4, Issue 2.
Improved algorithms for splitting full matrix algebras (arXiv).
Gábor Ivanyos, Ádám D. Lelkes, and Lajos Rónyai.
JP Journal of Algebra, Number Theory and Applications 28 (2013), 141--156.
Workshop Papers
Fair boosting: A case study (pdf). Benjamin Fish, Jeremy Kun, and Ádám D. Lelkes.
ICML 2015 Workshop on Fairness, Accountability, and Transparency in Machine Learning
(FAT ML 2015).
Biclique coverings, rectifier networks and the cost of ε-removal (arXiv).
Szabolcs Iván, Ádám D. Lelkes, Judit Nagy-György, Balázs Szörényi, and György Turán.
In Proceedings of the 16th International Workshop on Descriptional Complexity of Formal Systems
(DFCS 2014).
Patents
Crossword puzzle generator. US Patent 10,967,248
Control method for a cooling system with variable cooling power and cooling system. European Patent Office, EP17155305.0
Control method for an electrically excited motor and inverter. United States Patent and Trademark Office, US 15/413,433
Control method for a converter and converter. German Patent and Trademark Office, DE 10 2016 004 282.6
Control method for a cooling system with variable cooling power and cooling system. German Patent and Trademark Office, DE 10 2016 001 824.0
Control method for an electromagnetic motor and inverter. German Patent and Trademark Office, DE 10 2016 000 743.5
Selected Awards
Dean's Scholar Fellowship (2016)
"The Dean's Scholar Fellowship is a one-year, non-renewable award presented by the Dean of the Graduate College in recognition of a student's scholarly achievement. The award is intended to provide the most distinguished, advanced-level graduate students with a period of time dedicated solely to the completion of their programs."
Chicago Consular Corps Scholarship (2015)
"The Chicago Consular Corps Scholarship was created in 2007 as an annual scholarship by the Chicago Consular Corps, which is comprised of 79 international consular representative offices within Chicago. Nine scholarships are available for foreign students studying at UIC."
BUTE Faculty of Sciences, Scientific Student Conference, Discrete Mathematics Section, 1st Prize and Special Prize of the President (2011)
Professional activities
Program committee member for ICML 2020, ICML 2019, ICML 2018, ICML 2017
Conference reviewer for NeurIPS 2020, NeurIPS 2019, NIPS 2018, NIPS 2017, NIPS 2016, IJCAI 2016, ALT 2015