Department of Engineering Cybernetics
Norwegian University of Science and Technology
Elektro D/B2, D452C, Gløshaugen
Postdoc researcher working on automating and embedding methods in microscopy into mobile robotic platforms, for real-time targeted plankton-taxa identification and classification, that feeds the process of autonomous navigation control and mapping, which assists in modeling oceanographic phenomena. A new framework for this task is under development. Examples of ML methods explored are listed but not limited to: Mask-RCNN and attention segmentation to perform real-time segmentation; Learning the Bag-of-features to better classify identified objects; Active learning to minimize the labeling task effort; C-GAN to solve the class imbalance problem; and, methods of unsupervised learning to identify unknown extracted objects in-situ.
Robotic Vision, Computer Vision, Robust AI, Constraint Programming and Optimization, Reasoning about Uncertainty
Current Master thesis co-supervision:
- Unsupervised-Learning of the bag-of-features representation of planktonic organisms (Eivind Salvesen)
- Applying Deep-learning methods of One-shot / Zero-shot Image Recognition to discover new unknown classes (Andreas Langeland)
- Advanced techniques for segmentation to identify plankton-taxa and their distribution. (Jonas Borgersen)
- Applying active-learning techniques in machine learning to minimize the labeling effort (Martin Haug)
Previous Master thesis co-supervision (graduated June 2020)
- Applying Transfer Learning techniques to train deep learning over multiple datasets of planktons. (Sverre Anders Torp)
- Machine learning methodologies that solve the class-imbalance problem for datasets of planktonic organisms. (Oda Kiese)
- Advanced techniques for segmentation to identify plankton-taxa and their distribution. (Sondre Bergum)
Currently a member of the IDUN-ITK group at NTNU, which is part of the “IDUN – from Ph.D. to professor” project. The IDUN project main goal is to promote the female representation in academia. Thereof, the ITK group main research focus in a 2-years journey is to develop methodologies of Robust AI learning and to ensure their applicability to broad set of applications such as robotic control.
- Bergum, S., Saad, A., Stahl, A. (2020) Automatic in-situ instance and semantic segmentation of planktonic organisms using Mask R-CNN. IEEE Oceanic Engineering Society & Marine Technology Society - Singapore, 2020-10-5 - 2020-10-14.
- Salvesen, E., Saad, A., Stahl, A. (2020) Robust methods of unsupervised clustering to discover new planktonic species in-situ. IEEE Oceanic Engineering Society & Marine Technology Society - Singapore, 2020-10-5 - 2020-10-14.
- Teigen, A., Saad, A., Stahl, A. (2020) Leveraging Similarity Metrics to In-Situ Discover Planktonic Interspecies Variations or Mutations. IEEE Oceanic Engineering Society & Marine Technology Society - Singapore, 2020-10-5 - 2020-10-14.
- Ansari, S., Saad, A., Stahl, A., Rajachandran, M. (2020) Vision-based Real-time Zooplankton Detection and Classification using Faster R-CNN. Ocean Sciences Meeting 2020. AGU, ASLO and TOS; San Diego, CA. 2020-02-16 - 2020-02-21.
- Kiese, O., Saad, A., Stahl, A. (2020) Towards a Balanced-Labeled-Dataset of Planktons for a Better In-Situ Taxa Identification. Ocean Sciences Meeting 2020. AGU, ASLO and TOS; San Diego, CA. 2020-02-16 - 2020-02-21.
- Saad, A., Davies, E., Stahl, A. (2020) Recent Advances in Visual Sensing and Machine Learning Techniques for in-situ Plankton-taxa Classification. Ocean Sciences Meeting 2020. AGU, ASLO and TOS; San Diego, CA. 2020-02-16 - 2020-02-21.
- Saad, A., Stahl, A. (2019) AILARON - Autonomous Imaging and Learning Ai RObot identifying plaNkton taxa in-situ. Geilo Winter School 2019: Learning from Data. SINTEF; Geilo. 2019-01-20 - 2019-01-25.
- Saad, A. 2016. CDF-intervals: A Probabilistic Interval Constraint Framework to Reason about Data with Uncertainty. Open Access Repositorium der Universität Ulm. Dissertation. http://dx.doi.org/10.18725/OPARU-3966
- Saad, A., Frühwirth, T., and Gervet, C. 2014. The P-Box CDF-Intervals: A Reliable Constraint Reasoning with Quantifiable Information. The 30th International Conference on Logic Programming, ICLP2014, published in Theory and Practice of Logic Programming. pp. 461-475. DOI:10.1017/s1471068414000143
- Saad, A 2014. CDF-Intervals: A Reliable Framework to Reason about Data with Uncertainty. The 10th ICLP Doctoral Consortium.
- Saad, A., Gervet, C., and Frühwirth, T. 2012. CDF-Intervals Revisited. The Eleventh International Workshop on Constraint Modelling and Reformulation - ModRef2012.
- Saad, A., Gervet, C., and Frühwirth, T. 2012. CDF-Intervals: Reliable Constraint Reasoning with Quantifiable Information. The Proceeding of the Doctoral Program of the 2012 International Conference on Principles and Practice of Constraint Programming.
- Saad, A., Gervet, C., and Abdennadher, S. 2010. Constraint Reasoning with Uncertain Data Using CDF-Intervals. International Conference on the Integration of AI and OR Techniques in Constraint Programming, CPAIOR 2010, 292–306, Springer Verlag. DOI: 10.1007/978-3-642-13520-0_32s
- Mudawwar, M. and Saad, A. 2001. The k-ary n-cube Network and its Dual: a Comparative Study, in Proceedings of the 13th IASTED International Conference on Parallel and Distributed Computing and Systems, August 21-24, 2001, Anaheim, California, pages 254- 259.
- Presentation: From a Generic to a Customized Framework: Paving the Way for WebCT. Presented in Syllabus Fall 2002 Conference. November 2002, Boston Mariott Newton, Boston MA