Research
My research interests have mostly been in the domain of network science and data mining. In specific, I have been looking into the evolutionary dynamics of temporal networks like citation, human communication and author collaboration. More recently, I have been interested in applying deep neural architectures as well as social science theories to understand human behavior on the Web. I am also working towards developing methods for interpretable, fair and robust Machine Learning.
Publications:
Seham Nasr and Sandipan Sikdar. IndMask: Inductive Explanation for Multivariate Time Series Black-box Models. ECAI 2024, Santiago de Compostela, Spain
Maryam Badar, Sandipan Sikdar, Wolfgang Nejdl and Marco Fisichella. TrustFed: Navigating Trade-offs between Performance, Fairness, and Privacy in Federated Learning. ECAI 2024, Santiago de Compostela, Spain
Tobias Schumacher, Marlene Lutz, Sandipan Sikdar and Markus Strohmaier. Properties of Group Fairness Measures for Rankings. ACM Transactions on Social Computing. 2024
Jingge Xiao, Leonie Basso, Wolfgang Nejdl, Niloy Ganguly and Sandipan Sikdar. IVP-VAE: Modeling EHR Time Series with Initial Value Problem Solvers. AAAI 2024, Vancouver, Canada. Also presented at ICLR 2023 Workshop on Time Series Representation Learning for Health
Maryam Badar, Sandipan Sikdar, Wolfgang Nejdl and Marco Fisichella. FairTrade: Achieving Pareto-Optimal Trade-offs Between Balanced Accuracy and Fairness in Federated Learning. AAAI 2024, Vancouver, Canada
Sandipan Sikdar and Parantapa Bhattacharya. Interpretability of Deep Neural Networks. Ethics in Artificial Intelligence: Bias, Fairness and Beyond. Studies in Computational Intelligence, vol 1123. Springer, Singapore. 2023.
Manjish Pal, Subham Pokhriyal, Sandipan Sikdar and Niloy Ganguly. Ensuring Generalized Fairness in Batch Classification. Scientific Reports. 2023
Jan Engler*, Sandipan Sikdar*, Marlene Lutz and Markus Strohmaier (*equal contribution). SensePOLAR: Word sense aware interpretability for pre-trained contextual word embeddings. EMNLP (findings) 2022, Abu Dhabi, UAE. Also presented at BlackboxNLP 2022, IC2S2 2023 (Awarded one of the two best parallel talks)
Hussain Hussain*, Meng Cao*, Sandipan Sikdar, Denis Helic, Elisabeth Lex, Markus Strohmaier, and Roman Kern (*equal contribution). Adversarial Inter-Group Link Injection Degrades the Fairness of Graph Neural Networks. IEEE ICDM 2022, Orlando, Florida, USA
Jens Helge Reelfs, Timon Mohaupt, Sandipan Sikdar, Markus Strohmaier and Oliver Hohlfeld. Interpreting Emoji with Emoji: 🏖️ => ☀️😎🌊 . EMOJI@NAACL 2022, Seattle, Washington
Sandipan Sikdar, Florian Lemmerich and Markus Strohmaier. GetFair: Generalized Fairness Tuning of Classification Models. ACM FAccT 2022, Seoul, South Korea.
Sandipan Sikdar, Rachneet Singh Sachdeeva, Johannes Wachs, Florian Lemmerich and Markus Strohmaier. The Effects of Gender Signals and Performance in Online Product Reviews. Frontiers Big Data - Data Mining and Management - User Modeling and Recommendation (special issue) (fdata.2021.771404 ). 2021
Sandipan Sikdar*, Parantapa Bhattacharya* and Kieran Hesse (*equal contribution). Integrated Directional Gradients: Importance Attribution to Feature Interactions in Neural NLP Models. ACL 2021 (full paper), virtually (online).
Sandipan Sikdar, Animesh Mukherjee and Matteo Marsili. Unsupervised Ranking of Clustering Algorithms by INFOMAX. PLoS ONE (15 (10) e0239331). 2020
Binny Mathew*, Sandipan Sikdar*, Florian Lemmerich and Markus Strohmaier (*equal contribution). POLAR: Polar opposites enable interpretability of pre-trained word embeddings. Web Conference 2020 (previously WWW), Taipei, Taiwan
Abhirut Gupta, Sandipan Sikdar, Prateeti Mohapatra and Niloy Ganguly. Topic Influence Graph Based Analysis of Seminal Papers. CODS-COMAD 2020, Hyderabad, India
Soumya Sarkar, Bhanu Prakash Reddy, Sandipan Sikdar and Animesh Mukherjee. StRE: Self Attentive Edit Quality Prediction in Wikipedia. ACL 2019, Florence, Italy
Soumya Sarkar, Sandipan Sikdar, Sanjukta Bhowmick and Animesh Mukherjee. Using Core-Periphery Structure to Predict High Centrality Nodes in Time-Varying Networks. Data Mining and Knowledge Discovery Journal 32.5 (2018): 1368-1396 (ECML-PKDD 2018 journal track).
Sandipan Sikdar, Tanmoy Chakraborty, Soumya Sarkar, Niloy Ganguly and Animesh Mukherjee. ComPAS: Community Preserving Sampling for Streaming Graphs. AAMAS 2018, Stockholm, Sweden.
Sandipan Sikdar, Paras Tehria, Matteo Marsili, Niloy Ganguly and Animesh Mukherjee. On the effectiveness of the scientific peer-review system: a case study of the Journal of High Energy Physics. International Journal on Digital Libraries (IJDL) 2018.
Sandipan Sikdar. Significance of scientific peer-review system: A case study of the Journal of High Energy Physics. IEEE TCDL Bulletin 2017 and also in JCDL PhD Forum 2017, Toronto, Ontario, Canada.
Sandipan Sikdar, Matteo Marsili, Niloy Ganguly and Animesh Mukherjee. Influence of Reviewer Interaction Network on Long-term Citations: A Case Study of the Scientific Peer-Review System of the Journal of High Energy Physics. JCDL 2017, Toronto, Ontario, Canada. (nominated for best student paper)
Marcin Bodych, Niloy Ganguly, Tyll Krueger, Animesh Mukherjee, Rainer Seigmund-Schultze and Sandipan Sikdar (alphabetical order). Threshold-based epidemic dynamics in systems with memory. EuroPhysics Letters (EPL) 2017, 116(4), 48004
Sandipan Sikdar, Matteo Marsili, Niloy Ganguly and Animesh Mukherjee. Anomalies in the peer-review system: A case study of the journal of High Energy Physics. CIKM 2016, Indianapolis, USA.
Sandipan Sikdar, Abhijnan Chakraborty, Anshit Choudhury, Gourav Kumar, S.Kumar, Abhijeet Patil, Niloy Ganguly and Animesh Mukherjee. Identifying and Characterizing Sleeping Beauties on YouTube, CSCW 2016 (poster), San Francisco, USA.
Sandipan Sikdar, Niloy Ganguly and Animesh Mukherjee. Time series analysis of temporal networks. European Physics Journal B (EPJB) topical issue on Temporal Network Theory and Applications 2016, 89(1).
Sandipan Sikdar, Marcin Bodych, Rajib Ranjan Maity, Biswajit Paria, Niloy Ganguly, Tyll Krueger and Animesh Mukherjee. On the broadcast of segmented messages in dynamic networks. NetSciCom 2015, HongKong.
Tanmoy Chakraborty, Sandipan Sikdar, Niloy Ganguly and Animesh Mukherjee. Citation Interactions among Computer Science Fields: A Quantitative Route to the Rise and Fall of Scientific Research. Social Network Analysis and Mining (SNAM) 2014.
Tanmoy Chakraborty, Sandipan Sikdar, Vihar Tammana, Niloy Ganguly and Animesh Mukherjee. Computer Science Fields as Ground-truth Communities: Their Impact, Rise and Fall. ASONAM 2013, Ontario, Canada. (nominated for best paper)