Thorsten Joachims
Thorsten Joachims moved to Cornell in 2001 after doctoral work at Dortmund and a postdoctoral role at GMD. His research connects learning theory with human-centered information systems. His work includes SVM-light for support-vector learning and SVM-struct for complex outputs such as trees and alignments. His work also includes Propensity SVM-Rank for learning rankings from biased, partially labeled data. POEM uses counterfactual risk minimization for batch learning from bandit feedback. Related projects study counterfactual evaluation, exposure fairness and control in interactive information systems.
- Now at
- Cornell University
- Works
- 334
- Citations
- 51,952
- h-index
- 68
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2001
Joining Cornell
Thorsten Joachims moved to Cornell in 2001 after doctoral work at Dortmund and a postdoctoral role at GMD. His research connects learning theory with human-centered information systems.
-
Research software
Structured outputs and ranking
His work includes SVM-light for support-vector learning and SVM-struct for complex outputs such as trees and alignments. His work also includes Propensity SVM-Rank for learning rankings from biased, partially labeled data.
-
Research program
Learning from logged feedback
POEM uses counterfactual risk minimization for batch learning from bandit feedback. Related projects study counterfactual evaluation, exposure fairness and control in interactive information systems.
Sources and credits
Photo: www.cs.cornell.edu; embedded metadata: EXIF:Artist = Jason Koski/Cornell University; IPTC:By-line = Jason Koski/Cornell University; IPTC:By-lineTitle = Multimedia Producer; IPTC:Credit = Jason Koski/Cornell University; XMP:creator = Jason Koski/Cornell University. License: Not stated by source. Notice in the file: © Cornell University. Source: cs.cornell.edu. Cropped to a square and resized. Ask us to remove or replace this photo.