Big Data 30 min read

Insights and Paper Summaries from KDD 2017 Conference

The article provides a comprehensive overview of KDD 2017, including acceptance statistics, best paper awards, Ant Group's contributions, detailed discussions on AB testing, graph mining, and selected research papers across data mining, machine learning, and anomaly detection, offering valuable insights for practitioners and researchers.

AntTech
AntTech
AntTech
Insights and Paper Summaries from KDD 2017 Conference

KDD 2017, the premier ACM conference on data mining, attracted over 2,000 participants but accepted only about 200 papers, resulting in an acceptance rate below 20%.

The report highlights acceptance numbers for the Applied Data Science Track (ADST) and Research Track (RT), noting a 40% acceptance share for ADST and emphasizing popular topics such as temporal data and graph algorithms, while AI remained the most attended theme.

Best paper awards included the Best Paper/Best Student Paper (Accelerating Innovation Through Analogy Mining), the Best Applied Paper (HinDroid: An Intelligent Android Malware Detection System Based on Structured Heterogeneous Information Network), and the Doctoral Dissertation Award (Local Modeling of Attributed Graphs).

Ant Group’s presence featured a keynote on the Kunpeng distributed learning platform, describing its parameter‑server architecture, fault‑tolerance mechanisms, graph‑based scheduling, and successful deployments in Alibaba’s Double‑11 shopping festival and Ant’s risk detection systems.

The article then delves into AB testing, explaining its causal inference foundations, industrial adoption, and recent research advances presented at the conference, including tutorials from Microsoft EXP and workshops on advertising experiments.

A dedicated Graph section surveys recent advances such as struc2vec, metapath2vec, Toeplitz Inverse Covariance‑Based Clustering (TICC), and various graph‑clustering and link‑prediction methods, emphasizing their relevance to large‑scale network analysis.

The Risk Control/Anti‑Money‑Laundering segment summarizes 13 research papers and 7 application papers, highlighting anomaly detection techniques, heterogeneous information network models, and the award‑winning Hindroid malware detection system.

Finally, selected paper digests cover topics ranging from robust deep autoencoders for anomaly detection to Benford’s Law‑based digit anomaly detection, providing concise overviews of methodologies, experimental results, and practical implications.

AB testingbig dataMachine LearningData MiningAnomaly DetectionKDDgraph mining
AntTech
Written by

AntTech

Technology is the core driver of Ant's future creation.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.