Christian Walter Peter Omlin

My early research focused on the theoretical foundations of knowledge representation in recurrent neural networks with extensions to fuzzy domains as well as hybrid methods for learning; there is a resurgence of interest in this topic in cognitive science and neuroscience and in the wider context of explainable AI.

My main research interest is in the modelling, analysis and prediction/classification of spatiotemporal patterns using deep machine learning techniques as drivers of applications within the wider settings of converging technologies for Industry 4.0 (diagnostics and prognostics of industrial infrastructures, renewable energies, manufacturing, critical infrastructure protection, safety & security, health, finance, health, education). The ambition for autonomy must be balanced with the need for trust in autonomous systems. There is an emphasis on explainable AI and its relationship to other key imperatives of future AI systems (security, privacy, ethics, and morality) as part of overall AI governance.

Some research highligts are as follows:

1.       We have recently introduced a novel, generic reinforcement learning (RL) paradigm - affinity-based RL (ab-RL) where agents learn strategies that are partially decoupled from reward functions.  It solves problems with an intrinsic global affinity for certain actions, i.e. it encourages agents to learn a desired behavior while maximizing rewards.affinity-based RL (ab-RL) aims at (1) a parsimonious method based on a regularization of the objective function with a distinct action distribution that encourages RL agents (“prototypes”) to globally choose preferred actions based on desired agent traits, (2) an inherent RL interpretability which overcomes the obfuscation of opaque RL models that rely on post-hoc explanation and interpretation, (3) creation of mixed strategy agents thru fuzzy time-variant compositions of prototypical policies with hierarchical RL, each interpretable by its action affinities, that are globally interpretable, and (4) the paradigm allows identification of other agents' strategies without any need for inverse RL.

 2.       The Data Quality Monitoring (DQM) is in place to spot and diagnose particle physics data problems as promptly as possible to avoid data loss in the CMS experiment of CERN. Several studies have proposed to leverage the DQM automation using machine learning algorithms. However, only a few efforts explored temporal characteristics to underpin system monitoring automation of the CMS detectors via anomaly detection models thus far. We have developed a time-aware deep learning model for anomaly detection, prediction, and diagnostics on the multidimensional spatial quantity of the DQM for the HCAL detector.

3.       Machine ethics has received increasing attention over the past few years because of the need to ensure safe and reliable artificial intelligence (AI). The two dominantly used theories in machine ethics are deontological and utilitarian ethics. Virtue ethics is often mentioned as an alternative ethical theory. While this interesting approach has certain advantages over popular ethical theories, little effort has been put into engineering artificial virtuous agents due to challenges in their formalization, codifiability, and the resolution of ethical dilemmas to train virtuous agents. We motivate the implementation of virtuous agents that play such role-playing games, and the examination of their decisions through a virtue ethical lens. The development of such agents and environments is a first step towards practically formalizing and demonstrating the value of virtue ethics in the development of ethical agents.

IKT109 - Principles of Artificial Intelligence

IKT203 - Algorithms and Data Structures 

IKT301 - Applied Algorithms

IKT446 - Urban Computing

IKT462 - Digital Health: Fundamental (guest lecturer)

IKT463 - Artificial Intelligence in Society

IKT590 Master's Thesis (contact me for topics)

Work Experience

2022- Honorary Professor, School of Computing, University of South Africa.

2018- Professor, Department of ICT, University of Agder.

2016- Thesis Supervisor, School of Computing, University of South Africa.

2015-2016 Professor, School of Computing, University of South Africa.

2012-2014 Professor, School of Computer Science, University of the Witwatersrand, South Africa.

2007-2012 Professor, Department of Computer Engineering, Middle East Technical University, Cyprus.

2007 Visiting Research Professor, Department of Computer Science, University of the Western Cape, South Africa.

2004-2007 Professor, School of Computing, Information & Mathematical Sciences, University of the South Pacific, Fiji Islands.

2003-2006 Extraordinary Professor, Department of Computer Science, University of the Western Cape, South Africa.

2004 Visiting Professor, Dalle Molle Institute for Artificial Intelligence (IDSI),. Lugano, Switzerland.

2001-2004 Professor, Department of Computer Science, University of the Western Cape, South Africa.

1998-2001 Senior Lecturer, Department of Computer Science, University of Stellenbosch, South Africa.

1996-1998 Owner / Consultant, Adaptive Computing Technologies, Troy, USA.

1991-1996 Visiting Scientist, NEC Research Institute, Princeton,. USA.

1987-1991 Graduate Assistant, Rensselaer Polytechnic Institute, Troy, USA.

1987 Software Engineer, Hasler AG (Ascom), Berne, Switzerland.

Leadership Experience

2012-2014 Head of Computer Science, University of the Witwatersrand, South Africa.

2004-2007 Departmental leadership in the design and planning of the new Japan-Pacific ICT Center at the University of the South Pacific. The center became operational in 2012.

2002-2004 Founding Director, Telkom / Cisco Center of Excellence for IP and Internet Computing, University of the Western Cape, South Africa.

2003 Organization of the South African DACST Innovation Fund Consortium "HearSEE: A Platform for Sign Language Interaction" 

2001-2004 Deputy Chairperson, Department of Computer Science, University of the Western Cape, South Africa.

2001-2004 Information and Communication Technologies research thrust leader for the South African National Research Foundation Program for Historically Black Universities at the University of the Western Cape, South Africa.

1.       Machine Learning Applications for the Hadron Calorimeter of the CMS Detection (2020 - 2023)

Advances in data science have played a crucial role in the pursuit of high energy physics and the ability of physicists to exploit the full potential of the Large Hadron Collider (LHC) at CERN. The last decade has seen an explosion of machine learning applications in particle physics, particularly in the area of particle and event identification and reconstruction.  However, there are other areas that could benefit greatly from advances in data science, particularly the methods by which physicists monitor the quality of data during data collection.  Since the quality of collected data hinges completely on the quality of the detector components at the time of data-taking, the rapid identification and resolution of any detector anomalies will result in a larger amount of collected data of the highest quality (M. Asres, C.W. Omlin (PI, UiA),  J. Dittmann, P. de Barbaro, (CERN)).

2.       AIMWind: Analytics for Wind Farm Asset Management (2021-2023)

Due to the global shift towards sustainable energy, Norway’s energy companies are preparing to embrace wind energy through large wind farm (WF) projects. Presently, the levelized cost of energy from the wind industry is not competitive to conventional energy sources. Most of the existing WFs in the EU are supported by governmental subsidies, which may not be extended into the future, making wind energy economically unsustainable. The economic prospects improve significantly by extending the production time through LTE or repowering WFs beyond the EOL. The knowledge generated through continuous health evaluation can be utilized to improve WF operations by optimising the dual objectives of maximal power production while ensuring minimal degradation. Such a continuous evaluation of WT and WF health that is accurate and also quick enough to modify operations requires the development of novel methods for processing terabyte-scale data produced by the WFs. This is only possible through data analytics. Therefore, this project aims at leveraging the power of artificial intelligence (AI) methods such as big data analytics and machine learning (ML) to achieve the goal. AIMWind project will develop analytics to estimate WT-specific damage accumulation over the lifetime of each WT in a WF. The approach integrates measurements such as supervisory control and data acquisition (SCADA) data, condition monitoring systems (CMS) data, maintenance and inspection logs and meteorological data to assess the machine health Furthermore, the WF health status can be used to design dual objective controls for extended life as well as high efficiency. (M.S. Mathew, C.W. Omlin (WP leader), S.T. Kandukuri, V.K. Huynh  M. K.G. Robbersmyr, (UiA), R. Schlanbusch (NORCE), J.W. van Wingerden, R. Ferrari (TU Delft), F. Vanni (DNV-GL)).

3.       Machine Learning for Ultra-Precision Process Control and Optimization (2021-2024)

Ultra-precision engineering deals with surfaces polished to nanometre accuracy.  Some applications of such finely polished surfaces include optical components found in satellites, telescopes, lasers, spectrometers and thermal imagers as well as turbine blades or joint & cranial implants to name a few. Even though much of the process is carried out by computer numerically controlled (CNC) machines, at present, fully-autonomous bespoke manufacturing of such fine parts is still impossible. The complexity of the process, imperfect determinism, surface inaccessibility during the polishing process and required extensive expert knowledge, all pose major challenges in ultra-precision manufacturing. Artificial intelligence, as part of Industry 4.0 and progressing digital transformation efforts, brings hope of enhanced productivity, improved quality, and cost reduction. With a growing market for ultra-precision surfaces, increasing customer’s requirements, and skilled manpower, AI seems a promising solution for addressing the issues currently hindering autonomous production.  In this research aims  to advance ultra-precision manufacturing technology and deepen our understanding of factors that influence polishing performance by employing machine learning. We will focus on several aspects of the process, namely surface roughness prediction, polishing convergence, remaining useful life estimation and optimization of polishing parameters. The research will use data captured during industrial polishing runs. (M. Darowski, C.W. Omlin (PI),  M.F. Aftab (UiA), international research/industrial partners (confidentiality agreement)).

4.       Responsible Personal Financial Advisor (2020 – 2023)

In a rapidly growing FinTech industry, machine learning is becoming ever more prevalent and smarter solutions are pushing the boundaries of the status quo. In recent years, the European Union has implemented strict and specific regulations that govern the use of data and models alike, specifically the “Right to Explanation” and the “Right to Non-Discrimination”. Given these regulations, the application of AI in the financial sector is largely dependent on a model’s ability to provide a satisfactory explanation. In addition to regulations, there is an ethical obligation to ensure that any system implemented in the financial sector is fair, secure and conforms to privacy standards. These factors, along with explainability and interpretability, are the building blocks of responsible AI. This project aims to develop a system of machine learning models that provide financial advice to private customers, while conforming to the guidelines of Responsible AI. It will employ advanced analytics and modelling in order to provide customers with a more personalized and relevant financial advisory service. (C. Maree (SR Bank), C.W. Omlin (PI, UiA)).

5.       Learning Multimodal Intermediate Video and Language Representations in Deep Networks for Descriptive Object Identification and Tracking in Urban Environments (2021-2024)

Deep Learning models have revolutionized multimodal representation learning tasks, by learning joint representations with a common strategy across different modalities on top of modality specific networks. Nonetheless, it is still challenging to learn an adequate association between data modalities. Multimodal representation faces challenges to combine from heterogenous sources. This study aims to investigate correlations and associations between video and text modalities to fill the heterogeneity gap and semantic gap as well as preserving semantic correlations of multimodal features in common latent space. Inspired by the emerging deep learning architectures, we aim to test encoder/decoder based architectures for visual text semantic features extraction where deep multimodal transformer networks are used to learn powerful visual representations to jointly encode text into video. Transformer architecture leverages encoding and modelling temporal information. The developed  algorithms will be tested on the  AI City Challenge datasets which have several video feeds from real traffic surveillance cameras in urban environments. (C.W. Omlin (PI) UiA), T. Sadiq (UiA), K. Franke (NORCICS, NTNU)).

6.       AI4CITIZEN: Responsible AI for Citizen Safety in Future Smart Cities (2021 – 2024)

According to human rights standards and practice for the police that are put forward by the United Nations, police organisations shall ensure compliance with democratically agreed laws, prevent crimes, respond appropriately to emergencies, and provide supporting services to citizens. With the digitalisation of our society and adoption of technologies for sabotages, organised crimes, and terroristic attacks, police organisations are becoming increasingly dependent on advanced technologies in order to deal with the amount and complexity of information that occurs during their day-to-day operations. Our research aims to study the appropriateness of using AI in police operations to fulfill their mandate, and to foster human rights. In order to explore the effect of the potential introduction of non-intrusive technologies, we will contribute with novel research in three areas: anonymization of video monitoring data, anonymous crowd monitoring, and contextual video object identification and search. Thus the secondary objective is to develop novel ICT solutions in realm of societal security by performing technical research to improve security, to protect privacy, avoid sociocultural biases, at the same time exploring in which way new tools can help achieve the public acceptance and reduce social inequality biases. Our project objective is to develop solutions that bring balance in these trade-offs, improve surveillance technologies’ social acceptance and reduce social discrimination. It will thus contribute towards a balanced and responsible socio-technical approach in smart-city policing, where citizens can trust law enforcement. (L. Øverlier (Norwegian University of Science and Technology), C.W. Omlin (WP leader), L. Jiao, UiA), M. Dorotic (Norwegian Business School), United Nations Interregional Crime and Justice Research Institute, Oslo Municipality, Oslo Police District, Hessen Polizei, Netherlands National Police Lab AI, UK National Police Chiefs’ Council)).

7.       E-Health: IoT and AI for Patient Monitoring (2020 – 2023)

Quality care and patient safety are of vital importance for clinical personnel and hospital management. Today, clinical staff monitor inpatients with mental health issues on a 24/7 basis, including monitoring to assess inpatients' emotional/mental conditions. The monitoring is interruptive for patients, particularly during sleep, and is often regarded by patients as disturbing and interference in their private lives. Such procedures are costly and inefficient in terms of both workhour usage and workload for involved professionals. The rapid development of Internet of Things (IoT) technologies in recent year is reshaping the definition of future services. IoT is nowadays being developed and implemented globally in almost every societal sector including healthcare and smart homes. In the envisaged scenario described above, an IoT network can be deployed to screen patients’ movements and biomarkers as well as sending alerts when necessary. Machine learning holds great promise as an ideal tool for daytime and nighttime behavior analysis and anomaly detection for psychiatric patients. (M. Dutt (Egde Consulting); C.W. Omlin (Co-PI), M. Goodwin, F.Y. Li (UiA); Sørlandet Hospital)).

8.       Dilemmas and Choices in Games: From Machine Learning to Artificial Virtues in Intelligent Agents (2021 – 2024)

There is an urgent need to for the development of safe AI and its ethical use; simultaneously, it is necessary to understand the design of moral agents, i.e. agents whose behavior is aligned with human morality in critical applications such as healthcare, public service, and law. Artificial morality is the process of embedding an artificial agent with morality, thus rendering it capable of handling ethical dilemmas in a humanly acceptable manner. AI is bound to come across intricate dilemmas which have several right answers and is forced to make compromises on some of its values. This work investigates virtues within artificial agents by training these agents to solve a role-playing game riddled with moral dilemmas. Using frameworks such as deep reinforcement learning to train agents, we analyse the decisions using state-of-the-art explainable techniques to explain and interpret them through a virtue ethical lens, to truly understand whether the agent has developed virtues. This way, a platform to test a variety of artificial agents is developed to compare AI algorithms in terms of their inherent virtues. We also explore the prediction of virtues of other agents using techniques such as inverse reinforcement learning and Bayesian inference, and finally, to train an agent to behave virtuously to achieve favourable outcomes, given the uncertainty surrounding of virtues of another agent. Overall, this work aims to develop artificial virtuous agents. (A. Vishnawath, C.W. Omlin (PI), E. Bøhn, O.C. Granmo (UiA)).

Manuscripts under Review

A. Vishwanath, C.W. Omlin, “Reward Is Not All You Need for Designing Virtuous Agents with Reinforcement Learning”, International Conference on Machine Learning (ICML), 2023.

H. Md. Sharif, J. Lei, C.W. Omlin, “Deep Crowd Anomaly Detection by Fusing Reconstruction and Prediction Networks", Applied Intelligence, 2023.

H. Md. Sharif, J. Lei, C.W. Omlin, “Deep Crowd Anomaly Detection: State-of-the-Art, Challenges, and Future Research Directions”, Artificial Intelligence Review, 2023.

C. Maree, C.W. Omlin, “Symbolic Explanation of Affinity-Based Reinforcement Leaning Agents with Markov Models”, Expert Systems with Applications, 2022.

M.S. Mathew, S.T. Kandukuri, C. W. Omlin, “Predicting the Capacity Factor of Wind Farms using Explainable Deep Neural Networks for Estimation of Age Related Performance Degradation”, Wind Europe 2023, 2022.

W. Bounouaa, M.F. Aftaba, C.W. Omlin, “Online Detrended Fluctuation Analysis and Improved Empirical Wavelet Transform  for Real-time Oscillations Detection in Industrial Control Loops”, Computer and Chemical Engineering, 2022.

M.W. Asres, L. Wang, D. Yu, P. Parygin, J. Dittmann, C.W. Omlin, “Spatio-Temporal Anomaly Detection with Graph Networks for Data Acquisition Monitoring of the Hadron Calorimeter”, Sensors, 2022 (revised).

Publications in Refereed Journals

W. Saeed, C.W. Omlin, “Explainable AI (XAI): A Systematic Meta-Survey of Current Challenges and Future Opportunities”, Knowledge-Based Systems, 2023,  https://doi.org/10.1016/j.knosys.2023.110273 (citations: 15).

M. Dutt, S. Redhu, M. Goodwin, C.W. Omlin, “SleepXAI: An Explainable Deep Learning Approach for Multi-class Sleep Stage Identification”, Applied Intelligence, p. 1-14, 2022, https://doi.org/10.1007/s10489-022-04357-8.

A. Vishwanath, E.D. Bohr, O.C. Granmo, C. Maree, C. W. Omlin, “Towards Artificial Virtuous Agents: Games, Dilemmas and Mache Learning”, AI & Ethics, Springer, p. 1-10, 2022,  https://doi.org/10.1007/s43681-022-00251-8.

W. Bounouaa, M.F. Aftaba, C.W. Omlin, “Controller Performance Monitoring: A Survey of Problems and a Review of Approaches from a Data-driven Perspective with a Focus on Oscillations Detection and Diagnosis”, Industrial & Engineering Chemistry Research, p. 17735–17765, 2022, https://doi.org/10.1021/acs.iecr.2c02785.

C. Maree, C.W. Omlin, “Reinforcement Learning with Intrinsic Affinity for Personalized Prosperity Management”, Digital Finance, p. 241-262, 2022, https://doi.org/10.1007/s42521-022-00068-4.

C. Maree, C.W. Omlin, “Can Interpretable Reinforcement Learning Manage Assets Your Way?”, AI Vol. 3, No. 2, p. 526-537, 2022, https://doi.org/10.3390/ai3020030.

C. Maree, C.W. Omlin, “Reinforcement Learning Your Way: Agent Characterization through Policy Regularization”, AI, Vol. 3, No. 2, p. 250-259, 2022, https://doi.org/10.3390/ai3020015.

E. Zimudzi, I. Sanders, N. Rollings, C.W. Omlin, “Remote Sensing of Mangroves using Unmanned Aerial Vehicles: Current State and Future Directions”, Journal of Spatial Science, p. 195-212, 2019. (citations: 10)

E. Zimudzi, I. Sanders, N. Rollings, C.W. Omlin, “Segmenting Mangrove Ecosystems Drone Images using SLIC Superpixels, Geocarto International, p. 164-1662, 2018. (13 citations)

R.C. Staudemeyer, C.W. Omlin, “Extracting Salient Features for Network Intrusion Detection using Machine Learning Methods,’ South African Computer Journal, Vol. 52, pp. 82-96, 2014 (citations: 35).

R. Chandra, M. Frean, M. Zhang and C. W. Omlin, "Encoding Subcomponents in Cooperative Coevolutionary Recurrent Neural Networks", Neurocomputing, Elsevier,  Vol. 74, pp. 3223-3234, 2011 (citations: 35).

R. Chandra, R. Knight, C.W. Omlin, “Renosterveld Conservation in South Africa: A Case Study for Handling Uncertainty in Knowledge-Based Neural Networks for Environmental Management”, Journal of Environmental Informatics, Vol 13, No. 1, pp. 56-65, 2009 (citations: 8).

A. Sharma, C.W. Omlin, “Performance Comparison of Particle Swarm Optimization with Traditional Clustering Algorithms Used in Self-Organizing Maps”,  International Journal of Computational Intelligence, Vol. 5, No. 1, 2009 (citations: 14)

R. Chandra, C. W. Omlin, “Evolutionary training of Hybrid Systems of Recurrent Neural Networks and Hidden Markov Models”, International Journal of Applied Mathematics and Computer Science, Vol 3, No. 3, pp. 127-133, 2007.

A. Sharma, C.W. Omlin, “Determining Cluster Boundaries using Particle Swarm Optimization”, Proceedings of the World Academy of Science, Engineering & Technology, pp. 250-254, 2006. ( citations: 8)

C. Scheffler, K.H. Scheffler, C.W. Omlin, “Articulated Tree Structure from Motion - A Matrix Factorization Approach”, Transactions of the South African Institute of Electrical Engineers, Special Issue on Pattern Recognition: Theory and Applications, 2005. (citations: 2)

R.S. Kroon, C.W. Omlin, “Getting to Grips with Support Vector Machines: Theory”, South African Statistical Journal, Vol. 28, No. 2, p. 93-114, 2004 (citations: 1).

R.S. Kroon, C.W. Omlin, “Getting to Grips with Support Vector Machines: Application”, South African Statistical Journal, Vol. 28, No. 2, p. 159-172, 2004 (citations: 1).

C.W. Omlin, S. Snyders, ``Inductive Bias in Knowledge-Based Neural Networks: Application to Magnetic Resonance Spectroscopy of Breast Tissues'', Artifical Intelligence in Medicine, Vol. 28, No.2, p121-140, 2003 (citations: 13).

A. Vahed, C.W. Omlin, ``A Machine Learning Approach to Extraction of Symbolic Knowledge from Recurrent Neural Networks'', Neural Computation, Vol. 16, No. 1, p. 59-71, 2004 (citations: 14).

C.L. Giles, C.W. Omlin, K.K. Thornber, ``Equivalence in Knowledge Representation: Automata, Recurrent Neural Networks, and Dynamical Fuzzy Systems'', Proceedings of the IEEE (Special Issue on Computational Intelligence), D.B. Fogel, T. Fukuda, L. Guan (Eds.) , Vol. 87, No. 9, p. 1623-1640, 1999 (citations: 115).

C.W. Omlin, K.K. Thornber, C.L. Giles, ``Fuzzy Finite State Automata can be Deterministically Encoded into Recurrent Neural Networks'', IEEE Transactions on Fuzzy Systems, Vol. 6, No. 1, p. 76-89, 1998 (citations: 140).

C.W. Omlin, C.L. Giles, ``Constructing Deterministic Finite-State Automata in Recurrent Neural Networks'', Journal of the Association or Computing Machinery, Vol. 43, No. 6, p. 937-972, 1996 (citations: 237).

C.W. Omlin, C.L. Giles, ``Extraction of Rules from Discrete-Time Recurrent Neural Networks'', Neural Networks, Vol. 9, No. 1, p. 41-52, 1996 (citations: 268).

C.W. Omlin, C.L. Giles, ``Stable Encoding of Large Finite-State Automata in Recurrent Neural Networks with Sigmoid Discriminants'', Neural Computation, Vol. 8, No. 4, 1996 (citations: 69).

C.W. Omlin, C.L. Giles, ``Rule Revision with Recurrent Neural Networks'', IEEE Transactions on Knowledge and Data Engineering, Vol. 8, No. 1, p. 183-188, 1996 (citations: 77).

C.L. Giles, C.W. Omlin, ``Pruning Recurrent Neural Networks for Improved Generalization Performance'', IEEE Transaction on Neural Networks, vol. 5, No.5, p. 848-851, 1994 (citations: 145).

C.L. Giles, C.W. Omlin, ``Extraction, Insertion and Refinement of Symbolic Rules in Dynamically-Driven Recurrent Neural Networks'', Connection Science, vol. 5, no. 3-4, p. 307-337, 1993 (citations: 116).

Publications in Book Chapters

B. Wiese, C.W. Omlin, “Credit Card Transactions, Fraud Detection and Machine Learning: Modelling Time with LSTM Recurrent Neural Networks”, M. Bianchini, L. Jain, M. Maggini, F. Scraselli (Eds.),  Innovations in Neural Information Paradigms and Applications, Springer, pp. 231-268, 2009. (citations: 75).

P. Frasconi, C.L. Giles, M. Gori, and C.W. Omlin, ``Insertion of Prior Knowledge'', in J.F. Kolen and S.C. Kremer, (Eds.), A Field Guide to Recurrent Neural Networks, IEEE Press, pp. 155-177, 2001. (citations: 2).

C.W. Omlin, ``Understanding and Explaining DRN Behavior'', in J.F. Kolen and S.C. Kremer, (Eds.), A Field Guide to Recurrent Neural Networks, IEEE Press, p. 207-227, 2001. 

C.L. Giles, C.W. Omlin, ``Representation of Discrete States'', in J.F. Kolen and S.C. Kremer, (Eds.), A Field Guide to Recurrent Neural Networks, IEEE Press, pp. 83-102, 2001 (citations: 2).

C.W. Omlin, C.L. Giles, ``Symbolic Knowledge Representation in Recurrent Neural Networks: Insights from Theoretical Models of Computation'', in I. Cloete & J Zurada (Eds.), Knowledge-Based Neurocomputing, MIT Press, p. 63-105, 2000 (citations: 33).

C.W. Omlin, C.L. Giles, and K.K. Thornber, `` Fuzzy Knowledge and Recurrent Neural Networks: A Hybrid Dynamical Systems Perspective'', S. Wermter (Ed.), Hybrid Systems, Springer Verlag, 2000 (citations: 5).

C.W. Omlin, C.L. Giles, ``Extraction and Insertion of Symbolic Information in Recurrent Neural Networks'', V. Honavar, L. Uhr (Eds.), Artificial Intelligence and Neural Networks: Steps Toward Principled Integration, p. 271-299, Academic Press, 1994 (citations: 32).

Publications in International Refereed Conference Proceedings

M.S. Mathew,  S.T. Kandukuri, C.W. Omlin, “Estimation of Wind Turbine Performance DegradationWith Deep Neural Networks”, PHM Europe, p. 351-359, 2022.

utt, M. Goodwin, C.W. Omlin, “Sleep State Identification Based on Single-Channel EEG Signalsusing 1-D Convolutional Autoencoders”, IEEE Healthcom , 2022, accepted.

M. Asres, G. Cummings, P. Parygin, A. Khukhunaishvili, M. Toms, A. Campbell, S. I. Cooper, D. Yu, J. Dittmann, C. W. Omlin, “Unsupervised Multistep Anomaly Prediction using Causal Convolutional Autoencoders for Multivariate Time Series”, PHM Europe, p. 364-371, 2022.

C. Maree. C.W. Omlin, “Understanding Spending Behavior: Recurrent Neural Network Explanation and Interpretation”, IEEE Symposium on Computational Intelligence for Financial Engineering and Economics, 2022.

C. Maree. C.W. Omlin, “Balancing Profit, Risk, and Sustainability for Portfolio Management”, IEEE Symposium on Computational Intelligence for Financial Engineering and Economics, 2022.

M.W. Asres, G. Cummings, P. Parygin, A. Khukhunaishvili, M. Toms, A. Campbell, S. I. Cooper, D. Yu, J. Dittmann, C. W. Omlin, “Unsupervised Deep Variational Model for Multivariate Sensors Anomaly Detection”, International Conference on Progress in Informatics and Computing, p. 364-371,2021.

C. Maree, C.W. Omlin, “Clustering in Recurrent Neural Networks for Micro-Segmentation using Spending Personality”, IEEE Symposium Series on Computational Intelligence, 2021. (citations: 7)

M. Dutt, M. Goodwin, C.W. Omlin, “Automatic Sleep Stage Identification with Time Distributed Convolutional Neural Networks”, International Joint Conference on Neural Networks, 2021 (accepted).

C. Maree, J.E. Modal, C.W. Omlin, “Towards Responsible AI for Financial Transactions”, IEEE Symposium Series on Computational Intelligence”, p. 16-21, 2020. (citations: 8)

M. Holen, R. Saha, M. Goodwin, C.W. Omlin, K.E. Sandmark, “Road Detection for Reinforcement Learning Based Autonomous Car”, 3rd International Conference Information Systems, 2020. (citations: 6)

M. Holen, R. Saha, M. Goodwin, C.W. Omlin, “Road Detection for Reinforcement Based Autonomous Cars, International Conference on Information Science and Systems, ACM, p. 67-71, 2020.

R. Chandra, R. Deo, C.W. Omlin, “An Architecture for Encoding Two-Dimensional Cyclone Track Prediction in Coevolutionary Recurrent Neural Networks”, International Joint Conference on Neural Networks, p. 4865-4872, 2016. (citations: 5)

D. Bingol, T. Celik, C.W. Omlin, H.B. Vadapalli, “Facial Action Unit Intensity Estimating using Rotation-Invariant Features and Regressions Analysis”, International Conference on Image Processing, pp. 1381-1385, 2014 (citations: 9).

H. B. Vadapalli, H.O. Nyongesa, C.W. Omlin, “Classifying Facial Action Units: Use of Time Variant Data and Recurrent Neural Networks”,International Conference on Pattern Recognition and Information Processing, pp. 64-68, 2011.

H.B. Vadapalli, H. O. Nyongesa, C. W. Omlin. “Recurrent Neural Networks for Facial Action Unit Recognition from Image Sequences”, International Conference on Image Processing, Computer Vision, and Pattern Recognition, pp.357-361, 2009. (citations: 5)

R. Chandra, C. W. Omlin, ”Evolutionary One-Step Gradient Descent for Training Recurrent Neural Networks”, Proceedings of International Conference on Genetic and Evolutionary Methods, pp. 305-311, 2008.

R. Chandra, C.W. Omlin, “Combining Genetic and Gradient Descent Learning in Recurrent Neural Networks: An Application to Speech Phoneme Classification”, International Conference on Artificial Intelligence and Pattern Recognition, p. 278-285, 2007.

R. Chandra, C.W. Omlin, “A Hybrid Recurrent Neural Networks Architecture Inspired by Hidden Markov Models: Training and Extraction of Deterministic Finite Automaton”, International Conference on Artificial Intelligence and Pattern Recognition, p. 286-293, 2007.

R. Chandra and C. W. Omlin, “Hybrid Recurrent Neural Networks: An Application to Phoneme Classification” International Conference on Genetic and Evolutionary Methods, pp. 57-62, 2007 (citations: 2).

R. Chandra and C. W. Omlin, "Combining Genetic and Gradient Descent Learning in Recurrent Neural Networks: An Application to Speech Phoneme Classification", Proceedings of the International Conference on Artificial Intelligence and Pattern Recognition, pp. 278-285, 2007.

R. Chandra and C. W. Omlin, "Hybrid Recurrent Neural Networks Architecture Inspired by Hidden Markov Models: Training and Extraction of Deterministic Finite Automaton", Proceedings of the International Conference on Artificial Intelligence and Pattern Recognition,  pp. 286-293, 2007.

R. Chandra and C. W. Omlin, “Knowledge Discovery using Artificial Neural Networks for a Conservation Biology Domain” International Conference on Data Mining, pp. 221-227, 2007 (citations: 4).

D. Umuhoza, J. Agbinya, C.W. Omlin, “Estimation of Trust Metrics for MANET using QoS Parameter and Source Routing Algorithms”, 2nd International Conference on Wireless Broadband and Ultra Wideband Communications,  2007 (citations: 34).

R. Chandra, C.W. Omlin, “Training and Extraction of Fuzzy Finite State Automata in  Recurrent Neural Networks”, International Conference on Computational Intelligence, 2006, p. 274-279.

R. Chandra and C. W. Omlin, “Evolutionary Training of Hybrid Systems of Recurrent Neural Networks and Hidden Markov Models” Proceedings of the International Conference on Neural Networks, pp. 58-63, 2006.

A. Sharma, C.W. Omlin, “Cluster Extraction with Particle Swarm Optimimzation”, International Conference on Neural Networks, 2006.

D. Umuhoza, M. Momani, J. Agbinya, C.W. Omlin, ``On Trust Metric and Trust Modelling in Mobile Ad Hoc Networks'', IEEE ICTe Africa 2006, accepted for publication, 2006.

J. Whitehill, C.W. Omlin, ``Haar Features for FACS AU Recognition'', 7th International Conference on Automatic Face and Gesture Recognition (FGR06), pp. 97-101, 2006 (citations: 243) .

J. Whitehill, C.W. Omlin, ``Local versus Global Segmentation for Facial Expression Recognition'', 7th International Conference on Automatic Face and Gesture Recognition (FGR06), accepted for publication, pp. 357-362, 2006 (citations: 13).

O.P. Kogeda, J.I. Agbinya, C.W. Omlin, ``A probabilistic Approach to Faults Prediction in Cellular Networks'', Proceedings of the 5th IEEE International Conference on Networking  2006 (citations: 8).

O.A. Abidogun, C.W. Omlin,”Fraud Detection in Mobile Telecommunication Networks: Call Profiling with Unsupervised Neural Networks”, Proceedings of the 12th International IEEE Conference on Telecommunications, 2005.

R.C. Staudemeyer, D. Umuhoza,  C.W. Omlin, “Attacker Models, Traffic Analysis and Privacy Threats in IP Networks”, Proceedings of the 12th International IEEE Conference on Telecommunications, 2005. (citations: 5)

P. Kogeda, J.I. Agbinya, C.W. Omlin, ``Impacts and Cost of Faults on Services in Cellular Networks'', 4th International Conference on Mobile Business (ICMB2005), p. 551-555, 2005 (citations: 10).

R. Chandra, R. Knight, C.W. Omlin, ``A Neurocomputing Paradigm for Environmental Decision Support Systems'', International Conference on Environmental Management, 2005 (citations: 1).

R.K. Vadapalli, R. Knight, C.W. Omlin, ``Time Series Change Analysis of the Vegetation in Namaqualand, South Africa'',  International Conference on Environmental Management, 2005.

P. Kogeda , J.I. Agbinya and C.W. Omlin; “Probabilistic faults prediction in cellular networks”, 1st International Conference on Computers, Communications, and Signal Processing, 2005.

O.A. Abidogun, C.W. Omlin,”Call Profiling using Self-Organizing Maps”, in M. Mohammadian (Ed), Proceedings of the International Conference on Computational Intelligence for Modelling, Control and Automation, p. 371-379, 2004 (citations: 9).

D. Walsh, C.W. Omlin, ``Automatic Detection of Image Orientation with Support Vector Machines'', T. Hendtlass, M. Ali (Eds.), Developments in Applied Artificial Intelligence, LNAI, Vol. 2358, p. 36, 2002. Intelligence and Expert Systems, 2002.

J. van Zyl, C.W. Omlin, ``Knowledge-Based Neural Networks for Modelling Time Series'', International Workshop on Artificial Neural Networks, Lecture Notes in Computer Science, Springer Verlag, Vol. 2085, p. 579, 2001 (citations: 1).

J. van Zyl, C.W. Omlin, ``Prediction of Seismic Events in Mines using Neural Networks'', International Joint Conference on Neural Networks, p. 1492-1497, 2001 (citations: 6).

S. Snyders, C.W. Omlin, ``Inductive Bias in Recurrent Neural Networks'', Connectionist Models of Neurons, Learning Processes, and Artificial Intelligence Lecture Notes in Computer Science, Vol. 2084, Springer Verlag, p. 339-346, 2001 (citations: 3).

S. Snyders, C.W. Omlin, ``Rule Extraction from Knowledge-Based Neural Networks with Adaptive Inductive Bias'', International Conference on Neural Information Processing, Vol. 1, p. 143-148, 2001 (citations: 6).

A. Kruger, C.L. Giles, F. Coetzee, E. Glover, G.W. Flake, S. Lawrence, C. W. Omlin, ``DEADLINER: Building a Niche Search Engine'', Conference on Information and Knowledge Management, 2000 (citations: 50).

T. Wessels, C.W. Omlin, ``Refinement of Hidden Markov Models with Recurrent Neural Networks'', International Joint Conference on Neural Networks, 2000 (citations: §1).

T. Wessels, C.W. Omlin, ``A Hybrid System for Signature Verification'', International Joint Conference on Neural Networks, 2000 (citations: 75).

S. Snyders, C.W. Omlin, ``What Inductive Bias Gives Good Neural Network Training Performance?'', International Joint Conference on Neural Networks, 2000 (citations 19).

A. Vahed, C.W. Omlin, ``Rule Extraction from Recurrent Neural Networks using a Symbolic Machine Learning Algorithm'', Proceedings International Conference on Neural Information Processing, Vol. 2, p. 712-717, 1999 (citations: 26).

C.W. Omlin, K.K. Thornber, C.L. Giles, ``Deterministic Representation of Fuzzy Finite-State Automata in Recurrent Neural Networks'', Proceedings of the IEEE International Conference on Neural Networks, 1996 (citations: 11).

C.L. Giles, C. W. Omlin, B.G. Horne, ``Recurrent Neural Networks: Representation and Synthesis of Automata and Sequential Machines'' (invited paper), Proceedings of the IEEE International Conference on Neural Networks, 1996 (citations: 4).

C.L. Giles, C. W. Omlin, ``Learning, Representation, and Synthesis of Discrete Dynamical Systems in Continuous Recurrent Neural Networks'' (invited paper), Proceedings of the IEEE Workshop on Architectures for Semiotic Modeling and Situation Analysis in Large Complex Systems, J. Albus, A. Meystel, D. Pospelov, T. Reader (eds.), AdRem Inc., Balas Cynwyd, PA, p. 336, 1995. (citations: 6)

C.W. Omlin, C.L. Giles, B.G. Horne, L.R. Leerink, T. Lin, ``Training Recurrent Neural Networks with Temporal Input Encodings'', Proceedings of the IEEE International Conference on Neural Networks, (ICNN'94), p. 1267-1272, 1994 (citations: 1).

C.W. Omlin, C.L. Giles, ``Constructing Deterministic Finite-State Automata in Sparse Recurrent Neural Networks'', Proceedings of the IEEE International Conference on Neural Networks (ICNN'94), p. 1732-1737, 1994 (citations: 26).

C.W. Omlin, C.L. Giles, ``Integrating Temporal Symbolic Knowledge and Recurrent Neural Networks'' (invited paper), Proceedings of the International Symposium Integrating Knowledge and Neural Heuristics (ISIKNH'94), p.25-31, 1994.

C.W. Omlin, C.L. Giles, ``Pruning Recurrent Neural Networks for Improved Generalization Performance'', J. Vlontzos, J.N. Hwang, E. Wilson (Eds.), Neural Networks for Signal Processing IV: Proceedings of the 1994 IEEE Workshop, p. 690-699, 1994 (citations: 90)

C.W. Omlin, C.L. Giles, ``Rule Refinement with Recurrent Neural Networks'', Proceedings of the IEEE International Conference on Neural Networks (ICNN'93), Vol. 2, p. 801-806, 1993 (citations: 46).

C.W. Omlin, C.L. Giles, C.B. Miller, ``Heuristics for the Extraction of Rules from Discrete-Time Recurrent Neural Networks'', Proceedings of the International Joint Conference on Neural Networks (IJCNN'92), Baltimore, MD, Vol.1, p.33-38, 1992 (citations: 38).

C.L. Giles, C.W. Omlin, ``Inserting Rules into Recurrent Neural Networks'', S.Y. Kung,F. Fallside, J. A. Sorenson, C.A. Kamm (Eds.), Neural Networks for Signal Processing II: Proceedings of The 1992 IEEE Workshop, Copenhagen, Denmark, p.13-22, 1992 (citations: 49).

C.W. Omlin, C.L. Giles, ``Training Second-Order Recurrent Neural Networks using Hints'', D. Sleeman P. Edwards (Eds.), Proceedings of the Ninth International Conference on Machine Learning (ML'92), Morgan Kaufmann Publishers, San Mateo, CA, p.363-368, 1992 (citations: 82).

Publications in Local Refereed Conference Proceedings 

R.C. Staudemeyer, C.W. Omlin, “Evaluating Performance of Long Short-Term Memory Recurrent Neural Networks on Intrusion Detection Data”, Proceedings of the South African Institute for Computer Scientists and Information Technologist,  p. 218-224, 2013 (citations: 70).

H.B. Vadapalli, H. Nyongesa, and C.W. Omlin, "Recurrent Neural Networks for Facial Action Unit Recognition from Image Sequences", in Proceedings of the 21st Annual Symposium of the Pattern Recognition Association of
South Africa (PRASA), November, p.. 269-273, 2010.

R. Staudemeyer, C.W. Omlin, “Feature Set Reduction for Automatic Network Intrusion Detection with Machine Learning Algorithms”, Proceedings of the Southern Africa Telecommunication Networks and Applications Conference, 2009. (citations: 9)

O.P. Kogeda, J. Agbinya, C.W. Omlin, “Probabilistic Fault Prediction in Cellular Networks”, Proceedings of the Southern Africa Telecommunication Networks and Applications Conference, 2007 (citations: 18)

C. Abidogun, C.W. Omlin, “A Self-Organizing Maps Model for Outlier Detection in Call Data from Mobile Telecommunications Networks, Proceedings of the Southern Africa Telecommunication Networks and Applications Conference, 2004. (citations: 9)

O. Abidogun, C.W. Omlin, “Intelligent Fraud Detection”, Proceedings of the 1st African Conference on the Digital Common,, 2004 (citations: 1).

P. Kogeda, J. Agbinya, C.W. Omlin, “Faults and Service Modelling for Cellular Networks”, Proceedings of the Southern Africa Telecommunication Networks and Applications Conference, 2004. (citations: 6)

R.S. Kroon, C.W. Omlin, ``Bounding Generalization of Support Vector Machines", Proceedings of the 50th South African Statistical Conference, Johannesburg, South Africa, 2003.

C. Scheffler, K. Scheffler, C. W. Omlin, ``Articulated Tree Structure From Motion - A Matrix Factorisation Approach'', Proceedings of the fourteenth Annual Symposium of the Pattern Recognition Association of South Africa,  2003 (citations: 4).

C. Tiflin C.W. Omlin, “LSTM Recurrent Neural Networks for Signature Verification”, Proceedings of the Southern Africa Telecommunication Networks and Applications Conference, 2003. (citations: 12

M. Schulze, K. Scheffler, C.W. Omlin,  “Recognizing Facial Actions with Support Vector Machines”, Proceedings of the Thirteenth Annual Symposium of the Pattern Recognition Association of South Africa, 2002.

S. Naidoo, M. Glaser, C.W. Omlin, “Recognition of Static Hand Gestures for SASL Translation”, Proceedings of the 5th South African Conference on Telecommunications, Networks, and Applications, 2002 (citations: 22).

W.H. Andrag, C.W. Omlin, ``Q-Routing for Congestion Control in Telecommunication Networks with Limited Buffer Size'', Proceedings of the 4th South African Conference on Telecommunications, Networks, and Applications, 2001.

W.H. Andrag, C.W. Omlin, ``Optimization of Multiple Objectives in Telecommunication Networks using Intelligent Agents'', Proceedings 3rd South African Conference on Telecommunications, Networks, and Applications, p. 302-306, 2000.

W.H. Andrag, C.W. Omlin, ``Distributed Intelligent Multi-Agents for Telecommunication Network Management'', Proceedings 2nd South African Conference on Telecommunications, Networks, and Applications, p. 302-306, 1999.

T. Wessels, C.W. Omlin, ``Refining Hidden Markov Models with Recurrent Neural Networks'', Proceedings Proceedings 2nd South African Conference on Telecommunications, Networks, and Applications, p. 368-373, 1999 (citations: 11).

Unpublished Manuscript

O.C. Granmo, S. Glimsdal, L. Jiao, M. Goodwin, C.W. Omlin, G.T. Berg, “The Convolutional Tsetlin Machine”, 2019. (citations: 42)

J. Connan, C.W. Omlin, “Bibliography Extraction with Hidden Markov Models”, Department of Computer Science, University of Stellenbosch, 2000 (citations: 19).

F. Coetzee, A. Kruger, C.L. Giles, S. Lawrence, G. Flake, C.W. Omlin, “Binary Feature Selection and Integration in Specialized Search Engines, University of Stellenbosch, 2000.

C.W. Omlin, C.L. Giles, “Dynamic Adaptation of Recurrent Neural Network Architectures Guided by Symbolic Knowledge”, University of Stellenbosch, 1999.