Leman Akoglu, Heinz College (original) (raw)

Leman Akoglu

Dean's Associate Professor of Information Systems

2118C Hamburg Hall

412-268-3043

About me

Background

• Ph.D in Computer Science, Carnegie Mellon University, 2012 [thesis]
• B.S. in Computer Science, Bilkent University, 2003-2007

A short bio can be found here.

On academic leave

I am out-of-the-office on academic leave, as a full-time Amazon Scholar until Spring 2026.
I am hiring new PhD students to start in Fall 2026, however I may be delayed in my responses.

News:

• May 2025Our work on image anomaly detection with "self-tuning" self-supervision will appear at KDD 2025.

• Dec. 2024Our work on GPS trajectory anomaly detection and self-supervised time series anomaly detection are to appear at SIAM SDM 2025.

• Sep. 2024The preprint of our work on a foundation model for zero-shot outlier detection is now available.

• Sep. 2024Excited to start research and collaborations as a full-time Amazon Scholar during my first sabbatical year till Sep 2025!

• Sep. 2024Invited as a keynote speaker at International Conference on Automated Machine Learning (AutoML) 2024

• Sep. 2024Our work on continuous hyperparameter optimization for outlier models is to appear at AutoML 2024.

• Aug. 2024 Our work on automatic unsupervised hyperparameter tuning for deep models is to appear at ACM SIGKDD 2024 in Barcelona!

• Aug. 2024 Our work on algorithmic bias of outlier detection models is accepted at AAAI/ACM Conference on AI, Ethics and Society.

• July 2024Invited keynote speaker at the Graph Exploration (GraphEx) Symposium, Dedham, Massachusetts

• June 2024Invited speaker at the Fields Conference on Complex Networks in Banking and Finance, Toronto, Canada

• Oct. 2023 Our work on the pitfalls & opportunities of self-supervised learning for anomaly detection and multi-modal anomaly detection of attributed multi-graphs with metadata are to appear at IEEE BigData 2023.

• Aug. 2023 Our vision paper on the pitfalls and opportunities of self-supervised learning for anomaly detection is now on Arxiv.

• Aug. 2023 Excited to start research and collaborations as an Amazon Scholar!

• July 2023 Our research showcasing the role of augmentation on self-supervised anomaly detection will appear in TMLR 2023.

• Jul 2023 Speaker at the PFIA: French Platform on Artificial Intelligence, Strasbourg, France

• May 2023 Speaker at the Workshop on Modelling and Mining Complex Networks as Hypergraphs, Toronto Metropolitan University, Canada

• Dec 2022 Plenary speaker at the 2nd Online Winter School on Deep Learning: From Perceptrons to Transformers, Indian Statistical Institute (ISI) Kolkata, Feb 2023

• Dec 2022 Invited talk at NYU Stern School of Business Tech (IS) Research Seminar, Mar 1, 2023; NYC, USA

• Nov 2022 Pre-print of our study on Medicare fraud detection now available on arxiv.

• Sep. 2022 Our research on analysis and solutions to hyperparameter sensitivity of deep outlier models, quantifying expressiveness of graph neural networks, and graph-level anomaly detection are to appear at NeurIPS 2022. Congrats to all students and collaborators!

• Sep. 2022 Our research on unsupervised outlier model selection, clustered outlier detection, and graph-level anomaly detection are to appear at IEEE ICDM 2022. Congrats to all students!

• June 2022 Keynote speaker at the 9th International Conference on Information Management and Big Data (Nov. 16–18, 2022; Lima, Peru)

• June 2022 Keynote speaker at the 25th Discovery Science Conference (Oct. 10–12, 2022; Montpellier, France)

• June 2022 Our paper on summarizing node-labeled directed multi-graphs is to appear at ECML PKDD 2022.

• May 2022 Our paper on distributed outlier detection at scale (on Apache Spark) is to appear at KDD 2022 Applied Data Science track.

• Jan. 2022 Our paper on "lifting" the expressiveness of any GNN is to appear at ICLR 2022.

• Dec. 2021 Our ICDM 2021 paper on fast DOS-based graph embedding invited to KAIS Journal Special Issue (ICDM Best papers). Congrats Saurabh and Lingxiao!

• Oct. 2021 Talk at the Huawei Annual GlobalConnect Seminar

• Oct. 2021 Talk at the Google Scalable Algorithms for Semi-supervised and Unsupervised Learning Workshop

• Sep. 2021 Our paper on automating unsupervised outlier model selection is accepted to appear at NeurIPS 2021.

• Sep. 2021 Our survey paper on Graph Anomaly Detection with Deep Learning is accepted to appear in the IEEE TKDE Journal.

• Aug. 2021 I will give a keynote talk at the 30th ACM CIKM (International Conference on Information and Knowledge Management) in Nov 2021.

• Aug. 2021 Our paper on fast DOS-based graph embedding is to appear at IEEE ICDM 2021.

• Aug. 2021 Our survey paper on Graph Anomaly Detection using Deep Learning is now on Arxiv.

• May 2021 Selected as one of IJCAI 2021 Early-Career Spotlight speakers (among 15 all over the world)!

• Apr, 2021 Honored to receive the SDM/IBM Early Career Data Mining Research Award from the Society for Industrial and Applied Mathematics! This annual award recognizes one individual in the field of data science who has made outstanding, influential, and lasting contributions to the field within 10 years of having received a Ph.D.

• Apr, 2021 Joint work with Neil Shah @Snap on fairness-aware outlier detection to appear at AIES 2021, available on arXiv!

• Dec, 2020 Keynote speaker at the 9th International Conference on Complex Networks 2020 (Virtual), December 1-3. Slides are here. A similar version of the talk is on Youtube.

• Nov, 2020 OddBall (authored w/ Mary and Christos) received The Most Influential Paper Award at PAKDD 2020!

• Sep, 2020 Joint work with U Michigan on learning with graph neural networks under heterophily to appear at NeurIPS 2020!

• May, 2020 Our work on spotting emerging rare classes continually is to appear at ACM TKDD Journal!

• Apr, 2020 Couple of talks, Human-in-the-loop Anomaly Mining and Tackling Structure And Oversmoothing for Graph-based SSL, at the INFORMS Annual Meeting 2020.

• Apr, 2020 I am co-organizing the 3rd ACM SIGKDD 2020 Workshop on Machine Learning in Finance with Nitesh Chawla, Tanveer Faruquie, Senthil Kumar, Saurabh Nagrecha, and Jose A. Rodriguez-Serrano. Submissions here, due May 20th!

• Mar, 2020 I will attend the Amazon Scholar & Faculty Summit Fall 2020.

• Dec, 2019 Our work on tackling oversmoothing in graph neural networks via a novel normalization scheme to appear at ICLR 2020!

• Sep, 2019 Awarded an Adobe University Marketing Research Grant for our project titled_Entity Resolution and User De-duplication of Recurring Fraudsters and Free-loaders_. Thanks Adobe!

• Sep, 2019 Our work investigating nearest neighbor methods for anomaly detection to appear at NeurIPS 2019!

• Sep, 2019 I am giving an invited talk (highlighting the applications of my research on anomaly mining to fraud detection) at the AI in Finance Summit in NYC. See the agenda here.

• July, 2019 I am officially promoted to Associate Professor! Thanks to all the students, post-docs, collaborators, letter writers, and role models who helped make this happen :-)

• June, 2019 Our work on continual rare class classification to appear at ECML PKDD 2019!

• May, 2019 I am visiting and giving talks at Amazon Inc. at Palo Alto and the Stanford RAIN Seminar.

• Apr, 2019 I am honored to receive the prestigious title "Heinz College Dean's Professorship" for the next 3 years!

• Apr, 2019 Our work on human-in-the-loop interactive anomaly detection received the Best Research Paper Award at SIAM SDM'19!!!

• Mar, 2019 I am co-organizing the 2nd Workshop on Anomaly Detection in Finance at ACM SIGKDD 2019. Submit your related work by May 12!

• Feb, 2019 I am one of the Gilbreth Lecturers at the upcoming National Academy of Engineering's National Meeting! I will give a talk titled "Anomaly Mining: Detection and Beyond" to 100-200 middle and high school students in addition to the 40-50 NAE members in attendance. Here is a photo :-)

• Jan, 2019 Our work on contrastive and visual topic modeling to appear at TheWebConf'19 as a full paper!

• Dec, 2018 Our work on human-in-the-loop interactive anomaly detection to appear at SIAM SDM'19!

• Nov-Dec, 2018 I am visiting Singapore Management University for a collaborative project on early event detection using large-scale mobility data (bus, metro, autoparks, and social media).

• September, 2018 Our paper is awarded the Best Student Machine Learning Paper Runner-up at ECML PKDD 2018! The main scope of the work is to show why and how one could leverage privileged information for anomaly detection.

• August, 2018 2 conference papers on graph learning for semi-supervised classification and change point detection and localization on a graph accepted to appear at ACM CIKM'18!

• July, 2018 Our work on analyzing employee peer-reviews to appear at ACM/IEEE ASONAM'18 Industry Track!

• June, 2018 3 conference papers on anomaly detection and explanation accepted to appear at ECML PKDD'18!

• June, 2018 Our work on explaining anomalous patterns to appear at ECML PKDD'18 Journal Track!

• June, 2018 Upcoming talks and travel:

  • June 11: Talk titled A Quest for Structure: Joint Graph Structure and Semi-Supervised Inference [slides] at
    NetSci 2018 Satellite on Statistical Inference of Network Models (Paris, France)
  • June 12: Talk titled Attributed Networks: Social Circles, Summarization, Comparison [slides] at
    NetSci 2018 Satellite on Machine Learning In Network Science (Paris, France)
  • June 18-20: Talk and poster titled Anomaly Mining: Detection and Beyond at
    2018 Japan-American Frontiers of Engineering Symposium (Tsukuba, Japan)

• May, 2018 Our work on outlier detection in feature-evolving data streams to appear in ACM SIGKDD'18!

• Mar, 2018 I am co-organizing the 5th ACM SIGKDD 2018 Workshop ODD v5.0: Outlier Detection De-constructed,
  with Evgeny Burnaev (Skolkovo IST), Charu Aggarwal (IBM), Christos Faloutsos (CMU).
  Submit your ODD work by May 15th!

• Feb, 2018 Our work on graph-based semi-supervised learning in noisy multi-graphs to appear in PAKDD'18.

• Feb 9, 2018 Two talks at ACM WSDM: "Mining Rich Graphs: Ranking, Classification, and Anomaly Detection" at International Workshop on Heterogeneous Networks Analysis, and "Opinion Spam Detection: A Story of Networks, Meta-data and an Oracle" at MIS2: Workshop on Misinformation and Misbehavior Mining on the Web.

• Jan, 2018 Our work on attributed graphs: discovering communities and anomalies, interactive visual exploration and summarization appeared in Transactions on Knowledge Discovery from Data (TKDD) Journal, Volume 12, Issue 2.

• Nov, 2017 Talk at IEEE ICDM Workshop on High Performance Graph Data Mining and Machine Learning; Online Detection of Anomalous Heterogeneous Graphs with Streaming Edges

• Oct, 2017 Talk at UT Southwestern Medical Center CSB (Computational and Systems Biology) Seminar; autOmated Data Description (ODD): Explaining Anomalies for Human Interpretation

• Oct, 2017 Talk at UT Austin McCombs School of Business Information Management Seminar; Temporal Prediction of Customer Purchases and Using Forecasts in Coupon Design

• Sep, 2017 Talk at University of Pittsburgh Big Data Science Colloquium; Discovering Communities and Anomalies in Attributed Graphs: Interactive Visual Exploration and Summarization

• Aug, 2017 Talk at the ACM KDD Interactive Data Exploration and Analytics (IDEA) Workshop

• May, 2017 Our paper on promoting targeted time-limited digital coupons via purchase forecasts will appear at the Applied Data Science Track at ACM SIGKDD'17.

• Mar, 2017 Awarded an Adobe University Marketing Research Grant for our project titled_Real-time Detection of Online Click and Display Ad Exchange Fraud_. Thanks Adobe!

• Talk at the Women in Data Science (WinDS) Workshop at SIAM SDM, April 2017

• Jan, 2017 Our paper on spam URL detection to appear at PAKDD'17 and another work on explaining class differences via attributes to appear at WWW'17 Web Science Track.

• Dec, 2016 Our work on Ranking in Heterogeneous Networks with Geo-Location Information will appear at SIAM SDM'17.

• Sep, 2016 We are building ODDS (Outlier Detection Data Sets) -- an online data repository! Stay tuned for updates!

• Sep, 2016 1 (short) paper Sequential Ensemble Learning for Outlier Detection: A Bias-Variance Perspective is accepted to ICDM'16.

• Sep, 2016 Our DAMI article on a general framework for optimizing network robustness by edge rewiring is published in Volume 30, Issue 5.

• Aug, 2016 I moved to CMU'sHeinz College, where I will focus on data mining for societal problems through big data and computational methods. I am looking for students (PhD and MS)!

• July, 2016 Talk on anomaly mining at the US Army Research Labs.

• June, 2016 Talk on anomaly mining at the 2016 ICML Workshop on Anomaly Detection.

• June, 2016 Talk on opinion spam detection at Flipkart Inc. and Amazon India.

• May, 2016 1 paper Fast Memory-Efficient Anomaly Detection in Streaming Heterogeneous Graphs at ACM SIGKDD'16.

• May, 2016 Won SIAM SDM 2016 Best Paper Runner-up award for joint work with Bryan Perozzi on scalable ranking of anomalies in attributed graphs!

• Apr, 2016 Keynote speaker at the 12th International Workshop on Mining and Learning with Graphs (MLG) (co-located with KDD) on Aug 14th in San Francisco.

• Apr, 2016 Invited to serve as Workshops Co-Chair of ACM SIGKDD 2017.

• Apr, 2016 Invited speaker at the ICML 2016 Anomaly Detection Workshop on June 24th in NYC.

• Mar, 2016 Keynote speaker at the 2016 SDM Workshop on Mining Networks and Graphs: "Fraud Detection with Networks and Beyond" on May 7th in Miami.

• Mar, 2016 ACM SIGKDD 2016 Workshop ODD 4.0: Outlier Definition, Detection, and Description On-Demand, with F. Bell (Uber), E. Muller (HPI Germany), T. Senator (IARPA)

• Mar, 2016 1 short paper Temporal Opinion Spam Detection by Multivariate Indicative Signals to appear at ICWSM'16)

• Feb, 2016 Invited to serve as Workshops Co-Chair of SIAM SDM 2017.

• Feb, 2016 Our work on social security fraud detection with V. Van Vlasselaer, T. Eliassi-Rad, M. Snoeck, and B. Baesens is to appear in the Management Science Journal.

• Jan, 2016 Invited Associate Editorship for the IEEE Transactions on Knowledge and Data Engineering (TKDE) Journal.

• Jan, 2016 Invited to the Editorial Board of the Data Mining and Knowledge Discovery (DMKD) Journal.

• Jan, 2016 Our work on building anomaly ensembles is to appear in the Transactions on Knowledge Discovery from Data (TKDD) Journal.

• Dec, 2015 3 papers on attributed graph anomalies and opinion spam at SIAM SDM'16.

• Nov, 2015 1 paper Optimizing Network Robustness by Edge Rewiring: A General Framework at Data Mining and Knowledge Discovery (DAMI) (To appear at ECML PKDD'16)

• Nov, 2015 SIGKDD 2016 Student Travel Awards Co-chair

• Nov, 2015 ECML PKDD 2016 PhD Forum Co-chair

• Nov, 2015 Talk on "3D's of Anomaly Mining" at University of Illinois at Chicago

• Nov, 2015 Talk on "Semi-supervised Learning with Multi-Graphs" at Tepper School of Business at CMU

• Oct, 2015 Talk on "3D's of Anomaly Mining" at KU Leuven

• Oct, 2015 Attending Workshop on Information Networks (WIN) at NYU Business School (1 paper and 1 poster presentation).

• Sep, 2015 Invited speaker at the NII Shonan Meeting on "Analytics on Complex Networks: Scalable Solutions for Empirical Questions" Japan, Feb 2016.

• July, 2015 DARPA grant joint with IBM, Northwestern, UIC, and Stony Brook. Thanks DARPA!

• July, 2015 Offering a new course: Data Science Fundamentals this Fall.

• June, 2015 1 paper Discovering Opinion Spammer Groups by Network Footprints at ECML/PKDD'15.

• May, 2015 Received a Facebook Faculty Gift. Thanks Facebook!

• May, 2015 1 paper Collective Opinion Spam Detection: Bridging Review Networks and Metadata at ACM SIGKDD'15.

• Apr, 2015 Best Research Paper at SIAM SDM 2015. Congrats Hau and Shuchu!

• Apr, 2015 KDD 2015 tutorial "Graph-Based User Behavior Modeling: From Prediction to Fraud Detection" with Alex Beutel and Christos Faloutsos.

• Apr, 2015 Co-chairing KDD 2015 Workshop on Outlier Definition, Detection, and Description (ODDx3). Submit your work!

• Mar, 2015 NSF CAREER award (2015-2020). Thanks NSF! [Project_page]

• Feb, 2015 1 paper Correlation of Node Importance Measures: An Empirical Study through Graph Robustness at WWW'15 (Web Science Track).

• Jan, 2015 Invited talk at Workshop on Statistical and Computational Challenges in Networks, Web Mining, and Cybersecurity, Montreal, Canada (May 4-8).

• Dec, 2014 2 papers Less is More: Building Selective Anomaly Ensembles and Where Graph Topology Matters: The Robust Subgraph Problem at SIAM SDM'15.

• Nov, 2014 Invited talk at Huawei Technologies R and D, Santa Clara, CA.

• June, 2014 1 paper Guilt-by-Constellation: Fraud Detection by Suspicious Clique Memberships at HICSS'15.

• Sep, 2014 Slides and Software for our SIGKDD 2014 paper is now online.

• Sep, 2014 New course Data Mining meets Graph Mining this Fall.

• July, 2014 Invited talk at NICTA Sydney, AU

• July, 2014 Best paper award at ADC'14.

• June, 2014 1 paper Fast Nearest Neighbor Search on Large Time-Evolving Graphs at ECML/PKDD'14.

• June, 2014 2 papers Watch Your Tags: Analysis of Question Response Time in StackOverflow and Joint Voting Prediction for Questions and Answers in CQA at IEEE/ACM ASONAM'14.

• May, 2014 1 paper Focused Clustering and Outlier Detection in Large Attributed Graphs at ACM SIGKDD'14.

• April, 2014 DAMI Survey on Graph-based Anomaly Detection

• April, 2014 ODD^2 @ KDD2014: Workshop on Outlier Detection and Description under Data Diversity

• Mar 20, 2014 Attend Facebook Machine Learning Open House, Menlo Park, CA

• Mar 19, 2014 Invited talk at Twitter Inc., San Francisco, CA

• Mar 17, 2014 Invited talk at University of California, Santa Barbara

• Mar, 2014 1 (full) paper Quantifying Political Polarity based on Bipartite Opinion Networks is accepted to ICWSM'14.

• Mar, 2014 1 (full) paper ConnotationWordNet: Learning Connotation of the Word+Sense Network at ACL'14 for presentation.

• Feb, 2014 SDM Early Career Travel Grant, thanks SIAM!

• Jan, 2014 1 paper User Churn in Focused Q&A Sites: Characterizations and Prediction is accepted to WWW'14 WebSci for presentation.

• 1 (full) paper Make It or Break It: Manipulating Robustness in Large Networks is accepted to SIAM SDM'14.

• Apr 7, 2014 Full-day Workshop BGM @ WWW 2014 on Big Graph Mining

• 1 (short) paper on External Evaluation of Topic Models: A Graph Mining Approachis accepted to ICDM'13.

• Teaching Machine Learning this Spring!

• 1 paper on Sex Differences in the Human Connectome is accepted to BHI'13.

• Tutorial on Big Graph Mining at this year's ASONAM. Slides right here

• Check out my new course Networks and Data Mining Techniques

Research interests

• Outlier Mining, Out-of-Distribution Detection
• Fraud Detection, Event and Change Detection, Adversarial Example Detection
• Graph Neural Networks, Expressiveness, Graph Generation
• Graph-based/Relational Anomaly Detection
• Self-Supervised Learning
• Applied Machine Learning

Selected Honors and Awards

• IJCAI 2021 Early-Career Spotlight (among 15 all over the world)
• SDM/IBM Early Career Data Mining Research Award, April 2021
• The Most Influential Paper Award, PAKDD 2020
• Heinz College Dean's Professor for Feb 2019-2022
• Best Research Paper Award, SIAM SDM 2019
• Best Student Machine Learning Paper Runner-up Award, ECML PKDD 2018
• NSF CAREER Award, 2015-2020
• Best Research Paper Runner-up Award, SIAM SDM 2016
• Best Research Paper Award, SIAM SDM 2015
• Army Research Office Young Investigator Award, 2013
• Best Paper Award, PAKDD 2010
• Best Knowledge Discovery Paper Award, ECML PKDD 2009