Welcome!I am currently working as a Staff Research Scientist in Blue Prism AI Labs, London. Previously I was working as Principal Researcher (Machine Learning) in Bell Labs, Cambridge (UK).
I finished my Ph.D. in 2012 at the University of Leeds under Prof. Tony Cohn and Prof. David Hogg in the areas of Machine Learning and Cognitive Vision and was funded by the European Commission's Co-Friend, RACE, STRANDS projects, and DARPA's Mind's Eye project.
NewsThe AAAI-20 Workshop on Intelligent Process Automation
Staff Research Scientist
Principal Researcher, Machine Learning
Principal Researcher, Machine Learning
Member - Information Technology
PhD (Machine Learning)
Bachelor of Technology (Computer Science and Engineering)
2) Krishna S.R. Dubba; Miguel R. de Oliveira; Gi Hyun Lim; Hamidreza Kasaei; Luis Seabra Lopes; Ana Tome and Anthony G. Cohn: Grounding Language in Perception for Scene Conceptualization in Autonomous Robots. In AAAI Spring Symposium Series, 2014.
3) Rockel, S.; Neumann, B.; Zhang, J.; Dubba, K.; Cohn, A.; Konecny, S.; Mansouri, M.; Pecora, F.; Saffiotti, A.; Gunther, M.; Stock, S.; Hertzberg, J.; Tome, A.; Pinho, A.; Seabra Lopes, L.; von Riegen, S.; and Hotz, L: An Ontology-based Multi-Level Robot Architecture for Learning from Experiences. In AAAI Spring Symposium Series, 2013.
4) Krishna S.R. Dubba, Mehul Bhatt, Frank Dylla, Anthony G. Cohn, David C. Hogg: Interleaved Inductive-Abductive Reasoning for Learning Event-Based Activity Models, Proc. of ILP, 2011. Berkshire, UK.
6) Krishna S.R. Dubba, Subrat K. Dash, Arun K. Pujari: New Malicious Code Detection Using Variable Length n-grams, LNCS-4332/2006, Springer-Berlin / Heidelberg.
1) Subrat K. Dash, Krishna S.R. Dubba, Arun K. Pujari: New Malicious Code Detection Using Variable Length n-grams, Algorithms, Architectures and Information Systems Security, Statistical Science and Interdisciplinary Research - VOL3, World Scientific, 2008.
My work in general falls under the areas of Machine Learning (in particular Statistical Relational Learning, Deep Learning), Cognitive Vision, Knowledge Representation, Graph Analysis and Spatio-Temporal Reasoning. I worked in European Commission's RACE, STRANDS and Co-Friend projects. I also worked in DARPA's Mind's Eye project where we developed systems that learn relational event models from videos.
In order to behave autonomously, it is desirable for robots to have the ability to use human supervision and learn from different input sources (perception, gestures, verbal and textual descriptions etc). In this work, we present a cognitive architecture and learning framework for robot learning through natural human supervision and using multiple input sources by grounding language in perception.
One way to improve the robustness and flexibility of robot performance is to let the robot learn from its experiences. In this work, we propose an architecture and knowledge-representation framework for a service robot being developed in the EU project RACE. As a unique innovative feature, the framework combines memory records of low-level robot activities with ontology-based high-level semantic descriptions.
In this paper, we present a novel framework REMIND (Relational Event Model INDuction) for supervised relational learning of event models from large video datasets. The learned models can be used for recognizing events from previously unseen videos. The experimental results on several hours of video data from two challenging real world domains (an airport domain and a physical action verbs domain) suggest that the techniques are suitable to real world scenarios.
In this paper, we have proposed and successfully applied a novel supervised framework for learning relational event models from a huge, complex and noisy video dataset. The experimental results on video data from an airport apron where events such as Loading, Unloading, Jet-Bridge Parking etc. are learned suggests that the techniques are suitable to real world scenarios.
Motivated by the success of Machine Learning techniques in intrusion detection systems, recent research in detecting malicious executables is directed towards devising efficient non-signature based computer virus detection techniques that can profile the program characteristics from a set of training examples. In this paper, we describe a new feature selection measure, class-wise document frequency of byte n-grams. We empirically demonstrate that the proposed method is a better method for feature selection. For detection, we combine several classifiers using Dempster Shafer Theory for better classification accuracy instead of using a single classifier. Our experimental results show that such a scheme detects virus program far more efficiently than the earlier known methods.
Non-technical blog (Not updated anymore): Randomish
Floor 7, Aviation House,
125 Kingsway London,