Siavash Bahrami is a PhD candidate at Universiti Putra Malaysia (UPM), who is working on multimodal deep neural networks using acoustic and visual data for developing an active road safety system intended for autonomous and semi-autonomous vehicles. As part of the ULTRACEPT project working on work package 2, Siavash recently completed a 6 month secondment at the University of Lincoln and another at Visomorphic LTD.

Recently Siavash presented a paper titled “Acoustic Feature Analysis for Wet and Dry Road Surface Classification Using Two-stream CNN” during the 12th International Conference on Information and Multimedia Technology (ICIMT 2020). The data utilised for training and testing the proposed CNN architectures were collected during Siavash’s secondment in the UK. Despite the strains caused by the global pandemic Siavash managed to complete his secondment and collect all the data that was needed for his PhD thesis and the ULTRACEPT project with the help of UoL and UPM project members.

ICIMT 2020 was scheduled to take place between 11th and 13th December 2020 in Zhuhai, China, but due to the pandemic it was instead held as a virtual conference. The aim of ICIMT is to provide a platform for researchers, engineers, academics and industrial professionals from all over the world to present their research results and development activities in Information and Multimedia Technology. This conference provides opportunity for delegates to exchange new ideas and applications to establish business or research relations and to find global partners for future collaboration.

Professor Shigang Yue (ULTRACEPT Project Coordinator) and Associate Professor Dr. Shyamala Doraisamy (ULTRACEPT Project partner Lead for UPM) each chaired one of the conference sessions.

Acoustic Feature Analysis for Wet and Dry Road Surface Classification Using Two-stream CNN – Abstract

Road surface wetness affects road safety and is one of the main reasons for weather-related accidents. Study on road surface classification is not only vital for future driverless vehicles but also important to the development of current vehicle active safety systems. In recent years, studies on road surface wetness classification using acoustic signals have been on the rise. Detection of road surface wetness from acoustic signals involve analysis of signal changes over time and frequency-domain caused by interaction of the tyre and the wet road surface to determine the suitable features. In this paper, two single stream CNN architectures have been investigated. The first architecture uses MFCCs and the other uses temporal and spectral features as the input for road surface wetness detection. A two-stream CNN architecture that merges the MFCCs and spectral feature sets by concatenating the outputs of the two streams is proposed for further improving classification performance of road surface wetness detection. Acoustic signals of wet and dry road surface conditions were recorded with two microphones instrumented on two different cars in a controlled environment. Experimentation and comparative performance evaluations against single stream architectures and the two-stream architecture were performed. Results shows that the accuracy performance of the proposed two-stream CNN architecture is significantly higher compared to single stream CNN for road surface wetness detection.

Read more about Siavash’s ULTRACEPT work in his blog post here.

Yair Barnatan is an ULTRACEPT PhD student, at the University of Buenos Aires, working in the field of neuroethology. He is currently focused on neuronal processing of optic flow in the crustacean visual system, unravelling which and how neurons are involved in this process.

Yair attended the XXXV Annual Meeting of the Argentinian Society for Neuroscience Research, SAN 2020. This event was held virtually due to the global pandemic from 7th to 9th October, 2020.

This congress covered a wide variety of neuroscience topics, such as sensory and motor systems, neurodegenerative diseases and learning and memory. In that meeting, Yair presented a poster entitled “Functional evidence of the crustacean lobula plate as optic flow processing center” (Barnatan, Y., Tomsic, D. & Sztarker, J.)

Abstract

When an animal rotates it produces wide field image motion over its retina, termed optic flow (OF). OF blurs the image compromising the ability to see. Image shifts are stabilized by compensatory behaviors collectively termed optomotor response (OR). In most vertebrates and decapod crustaceans such reflex behavior involves mainly eye movements that consists in a slow tracking phase of the wide field image motion followed by a fast-resetting phase. We used the mud crab Neohelice granulata to tackle a major question in crustacean’s visual processing: which region of the brain is the neural substrate for processing OF? It has long been known that dipteran lobula plate (3rd optic neuropil) is the center involved in processing OF information. Recently, a crustacean lobula plate was characterized by neuroanatomical techniques, sharing many canonical features with the dipteran neuropil. In this work we present a functional evaluation of the role of crab’s lobula plate on the compensatory eye movements to rotational OF by performing electrolytic lesion experiments. We show that lesioning the lobula plate greatly impairs OR while keeping intact other visually guided behaviors, such as avoidance response upon an approaching stimulus. Even when OR is present in some lobula plate lesioned animals, these show reduced speed of eye tracking. Altogether, these results present strong evidence about an evolutionary conserved site for processing optic flow shared by crustacean and insects.

Nikolas Andreakos is a PhD candidate at the University of Lincoln, who is working on developing computational models of associative memory formation and recognition in the mammalian hippocampus.

Recently Nikolas attended the 13th International Conference on Brain Informatics (BI2020). Due to the current travel restrictions, this year’s conference, which was scheduled to take place on 19th September 2020 in Padova, Italy, was moved online.

About Brain Informatics 2020

The Brain Informatics (BI) conference series has established itself as the world’s premier research forum on Brain Informatics, which is an emerging interdisciplinary and multidisciplinary research field with joint efforts from neuroscience, cognitive science, medicine and life sciences, data science, artificial intelligence, neuroimaging technologies, and information and communication technologies.

The 13th International Conference on Brain Informatics (BI2020) provided a premier international forum to bring together researchers and practitioners from diverse fields for presentation of original research results, as well as exchange and dissemination of innovative and practical development experiences on Brain Informatics research, brain-inspired technologies and brain/mental health applications.

The theme of BI2020 was: Brain Informatics in the Virtual World.

The BI2020 solicits high-quality original research and application papers (both full paper and abstract submissions). Relevant topics included but were not limited to:

• Track 1: Cognitive and Computational Foundations of Brain Science
• Track 2: Human Information Processing Systems
• Track 3: Brain Big Data Analytics, Curation and Management
• Track 4: Informatics Paradigms for Brain and Mental Health Research
• Track 5: Brain-Machine Intelligence and Brain-Inspired Computing

Nikolas presented his research Andreakos N., Yue S., Cutsuridis V. (2020) Recall Performance Improvement in a Bio-Inspired Model of the Mammalian Hippocampus. In: Mahmud M., Vassanelli S., Kaiser M.S., Zhong N. (eds) Brain Informatics. BI 2020. Lecture Notes in Computer Science, vol 12241. Springer, Cham. https://doi.org/10.1007/978-3-030-59277-6_29.

Abstract

Mammalian hippocampus is involved in short-term formation of declarative memories. We employed a bio-inspired neural model of hippocampal CA1 region consisting of a zoo of excitatory and inhibitory cells. Cells’ firing was timed to a theta oscillation paced by two distinct neuronal populations exhibiting highly regular bursting activity, one tightly coupled to the trough and the other to the peak of theta. To systematically evaluate the model’s recall performance against number of stored patterns, overlaps and ‘active cells per pattern’, its cells were driven by a non-specific excitatory input to their dendrites. This excitatory input to model excitatory cells provided context and timing information for retrieval of previously stored memory patterns. Inhibition to excitatory cells’ dendrites acted as a non-specific global threshold machine that removed spurious activity during recall. Out of the three models tested, ‘model 1’ recall quality was excellent across all conditions. ‘Model 2’ recall was the worst. The number of ‘active cells per pattern’ had a massive effect on network recall quality regardless of how many patterns were stored in it. As ‘active cells per pattern’ decreased, network’s memory capacity increased, interference effects between stored patterns decreased, and recall quality improved. Key finding was that increased firing rate of an inhibitory cell inhibiting a network of excitatory cells has a better success at removing spurious activity at the network level and improving recall quality than increasing the synaptic strength of the same inhibitory cell inhibiting the same network of excitatory cells, while keeping its firing rate fixed.

When asked about his experience, Nikolas said:

“I really enjoyed the conference and learned a lot. It was a valuable and absorbing experience for me. The atmosphere was friendly. I shared my research and my experience with other attendants, and exchange ideas which would help me to improve my existing work”.

Nikolas Andreakos is a PhD candidate at the University of Lincoln who is working on developing computational models of associative memory formation and recognition in the mammalian hippocampus.

Recently, Nikolas attended the 9^th International Conference on Biomimetic and Biohybrid Systems, Living Machines 2020 (LM2020). This year’s conference which was scheduled to take place between 28^th and 30^th July 2020 in Freiburg, Germany, but due to the current situation around Covid-19, it was moved online.

Andreakos N., Yue S., Cutsuridis V. (2020) Improving Recall in an Associative Neural Network Model of the Hippocampus. In: Vouloutsi V., Mura A., Tauber F., Speck T., Prescott T.J., Verschure P.F.M.J. (eds) Biomimetic and Biohybrid Systems. Living Machines 2020. Lecture Notes in Computer Science, vol 12413. Springer, Cham. https://doi.org/10.1007/978-3-030-64313-3_1

ABOUT LIVING MACHINES 2020

The development of future real-world technologies will depend strongly on our understanding and harnessing of the principles underlying living systems and the flow of communication signals between living and artificial systems.

Biomimetics is the development of novel technologies through the distillation of principles from the study of biological systems. The investigation of biomimetic systems can serve two complementary goals. First, a suitably designed and configured biomimetic artefact can be used to test theories about the natural system of interest. Second, biomimetic technologies can provide useful, elegant and efficient solutions to unsolved challenges in science and engineering.

Biohybrid systems are formed by combining at least one biological component—an existing living system—and at least one artificial, newly-engineered component. By passing information in one or both directions, such a system forms a new hybrid bio-artificial entity. The theme of the conference also encompasses biomimetic methods for manufacture, repair and recycling inspired by natural processes such as reproduction, digestion, morphogenesis andmetamorphosis.

The following are some examples of “Living Machines” as featured at past conferences:

• Biomimetic robots and their component technologies (sensors, actuators, processors) that can intelligently interact with their environments.
• Biomimetic computers neuromimetic emulations of the physiological basis for intelligent behaviour.
• Active biomimetic materials and structures that self-organise and self-repair or show other bio-inspired functions.
• Nature inspired designs and manufacturing processes.
• Biohybrid brain-machine interfaces and neural implants.
• Artificial organs and body-parts including sensory organ-chip hybrids and intelligent prostheses.
• Organism-level biohybrids such as robot-animal or robot-human systems.

Nikolas presented his research Improving recall in an associative neural network model of the hippocampus.

Abstract

The mammalian hippocampus is involved in auto-association and hetero-association of declarative memories. We employed a bio-inspired neural model of hippocampal CA1 region to systematically evaluate its mean recall quality against different number of stored patterns, overlaps and active cells per pattern. Model consisted of excitatory (pyramidal cells) and four types of inhibitory cells: axo-axonic, basket, bistratified, and oriens lacunosum-moleculare cells. Cells were simplified compartmental models with complex ion channel dynamics. Cells’ firing was timed to a theta oscillation paced by two distinct neuronal populations exhibiting highly regular bursting activity, one tightly coupled to the trough and the other to the peak of theta. During recall excitatory input to network excitatory cells provided context and timing information for retrieval of previously stored memory patterns. Dendritic inhibition acted as a non-specific global threshold machine that removed spurious activity during recall. Simulations showed recall quality improved when the network’s memory capacity increased as the number of active cells per pattern decreased. Furthermore, increased firing rate of a presynaptic inhibitory threshold machine inhibiting a network of postsynaptic excitatory cells has a better success at removing spurious activity at the network level and improving recall quality than increased synaptic efficacy of the same threshold machine on the same network of excitatory cells, while keeping its firing rate fixed.

Fang Lei is a PhD Scholar at the University of Lincoln who is currently carrying out a 12 month ULTRACEPT secondment at project partner Guangzhou University, China. Fang is involved with Work Packages 1 & 2.

Fang Lei attended the IEEE World Congress on Computational Intelligence (WCCI) 2020. Originally due to take place in Glasgow between 19^th and 24^th July 2020, the conference was moved online due to Covid-19.

The WCCI is the world’s largest technical event on computational intelligence and features three conferences from the IEEE Computational Intelligence Society (CIS), the 2020 International Joint Conference on Neural Networks (IJCNN 2020), the 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2020), and the 2020 IEEE Conference on Evolutionary Computation (IEEE CEC 2020).

IEEE WCCI 2020 covered topics in the field of neural networks, from biological networks to artificial computation and was attended by more than 2,350 record attendees from over 75 countries. The event schedule included:

• Public lecture by Yoshua Bengio on the topic of artificial neural networks and deep learning 2.0
• 4 Plenary Speeches by world-renowned scholars: Barbara Hammer, Kay Chen Tan, Carlos Coello Coello, and Jim Bezdek
• 15 Keynotes by top-notch researchers, 5 per Conference
• 4 cutting-edge Panel sessions
• 36 Tutorials
• 10 Workshops
• 170 Special Sessions including 61 for IJCNN, 42 For IEEE CEC, 37 For FUZZ-IEEE, and 30 Cross-Disciplinary sessions
• 13 Challenging and contemporary competitions

Fang Lei presented her research Competition between ON and OFF Neural Pathways Enhancing Collision Selectivity which has been published by the IEEE:

F. Lei, Z. Peng, V. Cutsuridis, M. Liu, Y. Zhang and S. Yue, “Competition between ON and OFF Neural Pathways Enhancing Collision Selectivity,” 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, United Kingdom, 2020, pp. 1-8, doi: 10.1109/IJCNN48605.2020.9207131.

Abstract

The LGMD1 neuron of locusts shows strong looming-sensitive property for both light and dark objects. Although a few LGMD1 models have been proposed, they are not reliable to inhibit the translating motion under certain conditions compare to the biological LGMD1 in the locust. To address this issue, we propose a bio-plausible model to enhance the collision-selectivity by inhibiting the translating motion. The proposed model contains three parts, the retina to lamina layer for receiving luminance change signals, the lamina to medulla layer for extracting motion cues via ON and OFF pathways separately, the medulla to lobula layer for eliminating translational excitation with neural competition. We tested the model by synthetic stimuli and real physical stimuli. The experimental results demonstrate that the proposed LGMD1 model has a strong preference for objects in direct collision course-it can detect looming objects in different conditions while completely ignoring translating objects.

When asked about her experience of the conference, Fang Lei said…

“Although I hoped to attend the conference physically, I was still excited as it was my first time attending the international conference. It started at midnight in China due to the time difference but I was eager to share my research with academic peers and share this experience with them.

I was asked to present at the conference and did so via a pre-uploaded video. I presented within the visual system session of IJCNN regular sessions. The presentation went very smoothly and we discussed problems with presenting papers by asking and answering questions.

I found the conference to be an interesting and meaningful experience for me. It was good to be able to spread our work to peers and gain knowledge of the work others are doing. The only thing I wish was that the conference was face-to-face.”

University of Lincoln researcher Hamid Isakhani returned to China to continue his ULTRACEPT secondment at Huazhong University of Science and Technology in June of 2020. Due to the unprecedented Covid situation, Hamid was required to quarantine for 14 days in the Guangu Hotel. During this time, he virtually attended the Ubiquitous Robots (UR) 2020 Conference held in Kyoto, Japan.

Like many recent conferences, this was UR’s first ever virtual event. Organised by the Korea Robotics Society and co-sponsored by the IEEE Robotics and Automation Society, UR2020 brought together scientists and engineers across the world who are at the forefront of robotics and automation. The week-long conference held 22-26 June 2020 comprised a variety of Keynote speeches, workshops and break-out sessions. To read more about the conference you can visit the Ubiquitous Robots 2020 website, here.

Hamid presented his paper H. Isakhani, C. Xiong, S. Yue and W. Chen, “A Bioinspired Airfoil Optimization Technique Using Nash Genetic Algorithm,” 2020 17th International Conference on Ubiquitous Robots (UR), Kyoto, Japan, 2020, pp. 506-513, doi: 10.1109/UR49135.2020.9144868.

This paper was nominated for the URA 2020 Best Paper Award.

Here is what Hamid had to say about the experience:

“2020 has been no ordinary from day one. Plane crashes, earth quakes, forest fires and now the devastating pandemic has certainly changed all of our lives for good or bad (reader’s perception). What is important is that we learn, adapt, and live on. Although some of our decisions are better than the others; like the AERO 2020 conference in France being postponed by over a year due to the pandemic compelling us to withdraw our participation, KROS on the other hand decided to take the leap and conduct the 17th International Conference on Ubiquitous Robots (UR2020) in Kyoto, Japan on the originally scheduled dates virtually for the first time in its history. It was certainly challenging for both the organisers as well as the participants, yet it was a success with a lot of takeaways for everyone.

Although the core topic of our work is more related to the field of Aerospace, we were extremely pleased to learn that our work was greatly recognised by the robotics community and nominated for the best contributed paper award at UR2020.

On 25th June 2020, 09:00hrs (JST), our 10 minutes long pre-recorded presentation was played back on Zoom application for the audience who later raised their questions via the built-in Q&A tab provided on the platform. Participants had access to the audio and video of the presenter and the session Chair who communicated and sorted the posted questions through a one-on-one video call.

Online conference sceptics might argue that networking and physical meetings at an international conference is a significant advantage missing in a virtual event, especially for early career researchers. Fortunately, this feature was also thoughtfully integrated by the organisers on the Slack application where different channels were created for presenters to communicate and share opinions/contacts for a period of thirty days.

Overall, it was rather an interesting experience, although I was in my 14-days covid-19 quarantine in China, at least I didn’t have to attend my session past midnight in the UK (BST).”

Hamid remains in China to carry on his secondment activities, although not in quarantine anymore. He continues his study on the effects of haemolymph on the flexural stiffness of various flying insect species. You can learn more about Hamid’s research here.

Based on the successful mind-matching session at the Cognitive Computational Neuroscience (CCN) conference, a free web-based unconference for neuroscientists was created called “neuromatch“.

The neuromatch 1.0 conference was held on 30th and 31st March, 2020. The conference agenda included a significant number of international speakers.

Our ULTRACEPT researcher Xuelong Sun presented his work on insect navigation at the conference. Considering the current travel restrictions caused by Covid-19, this was an excellent opportunity to continue to promote the ULTRACEPT project work in an innovative, safe and effective way.

Xuelong presented his work ‘A Decentralised Neural Model Explaining Optimal Integration Of Navigational Strategies in Insects’. Xuelong is carrying out this work with Dr Michael Mangan and Prof Shigang Yue.

A copy of Xuelong’s presentation can be accessed here.

To learn more about this research, please refer to the paper Modelling the Insect Navigation Toolkit: How the Mushroom Bodies and Central Complex Coordinate Guidance Strategies https://doi.org/10.1101/856153 .

The IEEE 4th International Conference on Advanced Robotics and Mechatronics (ICARM) conference was held in Toyonaka Campus, Osaka University, Osaka, Japan on the 3rd to 5th July 2019.  You can access the conference website here.

The 2019 conference was collaboratively organized by robotic researchers from Osaka University, The University of Tokyo, Nara Institute of Science and Technology, and Ritsumeikan University, Japan. The conference provided an international forum for researchers, educators, engineers in general areas of mechatronics, robotics, automation and sensors to disseminate their latest research results and exchange views on the future research directions of these fields.

This conference was attending by ULTRACEPT researcher Tian Liu from the University of Lincoln. Tian presented the following research:

X. Sun, T. Liu, C. Hu, Q. Fu and S. Yue, “ColCOS Φ: A Multiple Pheromone Communication System for Swarm Robotics and Social Insects Research,” 2019 IEEE 4th International Conference on Advanced Robotics and Mechatronics (ICARM), Toyonaka, Japan, 2019, pp. 59-66, doi: 10.1109/ICARM.2019.8833989.

Abstract: In the last few decades we have witnessed how the pheromone of social insect has become a rich inspiration source of swarm robotics. By utilising the virtual pheromone in physical swarm robot system to coordinate individuals and realise direct/indirect inter-robot communications like the social insect, stigmergic behaviour has emerged. However, many studies only take one single pheromone into account in solving swarm problems, which is not the case in real insects. In the real social insect world, diverse behaviours, complex collective performances and flexible transition from one state to another are guided by different kinds of pheromones and their interactions. Therefore, whether multiple pheromone based strategy can inspire swarm robotics research, and inversely how the performances of swarm robots controlled by multiple pheromones bring inspirations to explain the social insects’ behaviours will become an interesting question. Thus, to provide a reliable system to undertake the multiple pheromone study, in this paper, we specifically proposed and realised a multiple pheromone communication system called ColCOSPhi. This system consists of a virtual pheromone sub-system wherein the multiple pheromone is represented by a colour image displayed on a screen, and the Colias IV micro-robots platform designed for swarm robotics applications. Two case studies are undertaken to verify the effectiveness of this system: one is the multiple pheromone based on an ant’s forage and another is the interactions of aggregation and alarm pheromones. The experimental results demonstrate the feasibility of ColCOSPhi and its great potential in directing swarm robotics and social insects research.

The 2019 International Joint Conference on Neural Networks (IJCNN) was held at the InterContinental Budapest Hotel in Budapest, Hungary on the 14-19 July 2019. The full Program with Abstracts can be found here.

This conference was attended by  ULTRACEPT researchers from the University of Lincoln, Huatian Wang and Hongxin Wang.

Neural Models of Perception, Cognition and Action

Tuesday, July 16, 5:30PM-7:30PM

Hongxin Wang presented the following:

Visual Cue Integration for Small Target Motion Detection in Natural Cluttered Backgrounds [#19188]

Hongxin Wang, Jigen Peng, Qinbing Fu, Huatian Wang and Shigang Yue, University of Lincoln, United Kingdom; Guangzhou University, China.  

The robust detection of small targets against cluttered background is important for future artificial visual systems in searching and tracking applications. The insects’ visual systems have demonstrated excellent ability to avoid predators, find prey or identify conspecifics – which always appear as small dim speckles in the visual field. Build a computational model of the insects’ visual pathways could provide effective solutions to detect small moving targets. Although a few visual system models have been proposed, they only make use of small-field visual features for motion detection and their detection results often contain a number of false positives. To address this issue, we develop a new visual system model for small target motion detection against cluttered moving backgrounds. Compared to the existing models, the small-field and wide-field visual features are separately extracted by two motion-sensitive neurons to detect small target motion and background motion. These two types of motion information are further integrated to filter out false positives. Extensive experiments showed that the proposed model can outperform the existing models in terms of detection rates.

Plenary Poster Session POS2: Poster Session 2

Thursday, July 18, 10:00AM-11:40AM

Huatian Wang presented the following:

P333 Angular Velocity Estimation of Image Motion Mimicking the Honeybee Tunnel Centring Behaviour [#19326]

Huatian Wang, Qinbing Fu, Hongxin Wang, Jigen Peng, Paul Baxter, Cheng Hu and Shigang Yue, University of Lincoln, United Kingdom; Guangzhou University, China

Insects use visual information to estimate angular velocity of retinal image motion, which determines a variety of flight behaviours including speed regulation, tunnel centring and visual navigation. For angular velocity estimation, honeybees show large spatial-independence against visual stimuli, whereas the previous models have not fulfilled such an ability. To address this issue, we propose a bio-plausible model for estimating the image motion velocity based on behavioural experiments of the honeybee flying through patterned tunnels. The proposed model contains mainly three parts, the texture estimation layer for spatial information extraction, the delay-and-correlate layer for temporal information extraction and the decoding layer for angular velocity estimation. This model produces responses that are largely independent of the spatial frequency in grating experiments. And the model has been implemented in a virtual bee for tunnel centring simulations. The results coincide with both electro-physiological neuron spike and behavioural path recordings, which indicates our proposed method provides a better explanation of the honeybee’s image motion detection mechanism guiding the tunnel centring behaviour.