Brains display very high-level parallel computation, fault-tolerance, and adaptability, all of which are what we struggle to recreate in engineered systems. The neurocomputer (an organic computer built from living neurons) seems possible and may lead to a new generation of computing device that can operate in a brain-like manner. Cultured neuronal networks on multi-electrode arrays (MEAs) are one of the best candidates for the neurocomputer for their controllability, accessibility, flexibility, and the ability to self-organize. I explored the possibility of the neurocomputer by studying whether we can show goal-directed learning, one of the most fascinating behavior of brains, in cultured networks. Inspired by the brain, which needs to be embodied in some way and interact with its surroundings in order to give a purpose to its activities, we have developed tools for closing the sensory-motor loop between a cultured network and a robot or an artificial animal (an animat), termed a ¡§hybrot¡¨. In order to efficiently find an effective closed-loop design among infinite potential options, I constructed a biologically-inspired simulated network. By using this simulated network, I designed: (1) a statistic that can effectively and efficiently decode network functional plasticity, and (2) feedback stimulations and an adaptive training algorithm to encode sensory information and to direct network plasticity. By closing the sensory-motor loop with these decoding and encoding designs, we successfully demonstrated a simple adaptive goal-directed behavior: learning to move in a user-defined direction, and further showed that multiple tasks could be learned simultaneously. These results suggest that even though a cultured network lacks the 3-D structure of the brain, it still can be functionally shaped and show meaningful behavior. To our knowledge, this is the first demonstration of goal-directed learning in embodied cultured networks. Extending from these findings, I further proposed a research plan to optimize closed-loop designs for evaluating the maximal learning capacity (or even true intelligence) of the cultured network. Knowledge gained from effective closed-loop designs provides insights about learning and memory in the nervous system, which could influence the design of neurocomputers, future artificial neural networks, and more effective neuroprosthetics.

Many biological systems are known to accomplish complex tasks in a decentralized, robust, and scalable manner - characteristics that are desirable to the coordination of engineered systems as well. Inspired by nature, we produce coordination strategies for a network of heterogenous agents and in particular, we focus on intelligent collective systems. Bottlenose dolphins and African lions are examples of intelligent collective systems since they exhibit sophisticated social behaviors and effortlessly transition between functionalities. Through preferred associations, specialized roles, and self-organization, these systems forage prey, form alliances, and maintain sustainable group sizes. In this thesis, we take a three-phased approach to bioinspiration: in the first phase, we produce agent-based models of specific social behaviors observed in nature. The goal of these models is to capture the underlying biological phenomenon, yet remain simple so that the models are amenable to analysis. In the second phase, we produce bio-inspired algorithms that are based on the simple biological models produced in the first phase. Moreover, these algorithms are developed in the context of specific coordination tasks, e.g., the multi-agent foraging task. In the final phase of this work, we tailor these algorithms to produce coordination strategies that are ready to be deployed in target applications.

Surveillance systems are usually employed to monitor wide areas in which their usersaim to detect and/or observe events or phenomena of their interest. The use ofwireless sensor networks in such systems is of particular interest as these networks can provide a relative low cost and robust solution to cover large areas. Emerging applications in this context are proposing the use of wireless sensor networks composed of both static and mobile sensor nodes. Motivation for this trend is toreduce deployment and operating costs, besides providing enhanced functionalities.The usage of both static and mobile sensor nodes can reduce the overall systemcosts, by making low-cost simple static sensors cooperate with more expensive andpowerful mobile ones. Mobile wireless sensor networks are also desired in somespecific scenarios in which mobility of sensor nodes is required, or there is a specificrestriction to the usage of static sensors, such as secrecy. Despite the motivation,systems that use different combinations of static and mobile sensor nodes are appearing and with them, challenges in their interoperation. This is specially the case for surveillance systems.This work focuses on the proposal of solutions for wireless sensor networks including static and mobile sensor nodes specifically regarding cooperative andcontext aware mission setup and performance. Orthogonally to the setup and performance problems and related cooperative and context aware solutions, the goalof this work is to keep the communication costs as low as possible in the executionof the proposed solutions. This concern comes from the fact that communication increases energy consumption, which is a particular issue for energy constrained sensor nodes often used in wireless sensor networks, especially if battery supplied. Inthe case of the mobile nodes, this energy constraint may not be valid, since their motion might need much more energy. For this type of node the problem incommunicating is related to the links’ instabilities and short time windows availableto receive and transmit data. Therefore, it is better to communicate as little as possible. For the interaction among static and mobile sensor nodes, all thesecommunication constraints have to be considered.For the interaction among static sensor nodes, the problems of dissemination and allocation of sensing missions are studied and a solution that explores local information is proposed and evaluated. This solution uses mobile software agentsthat have capabilities to take autonomous decisions about the mission dissemination and allocation using local context information so that the mission’s requirementscan be fulfilled. For mobile wireless sensor networks, the problem studied is how to perform the handover of missions among the nodes according to their movements.This problem assumes that each mission has to be done in a given area of interest. In addition, the nodes are assumed to move according to different movement patterns,passing through these areas. It is also assumed that they have no commitment in staying or moving to a specific area due to the mission that they are carrying. To handle this problem, a mobile agent approach is proposed in which the agents implement the sensing missions’ migration from node to node using geographical context information to decide about their migrations. For the networks combining static and mobile sensor nodes, the cooperation among them is approached by abiologically-inspired mechanism to deliver data from the static to the mobile nodes.The mechanism explores an analogy based on the behaviour of ants building and following trails to provide data delivery, inspired by the ant colony algorithm. It is used to request the displacement of mobile sensors to a given location according tothe need of more sophisticated sensing equipment/devices that they can provide, so that a mission can be accomplished.The proposed solutions are flexible, being able to be applied to different application domains, and less complex than many existing approaches. The simplicity of the solutions neither demands great computational efforts nor large amounts of memory space for data storage. Obtained experimental results provide evidence of the scalability of these proposed solutions, for example by evaluatingtheir cost in terms of communication, among other metrics of interest for eachsolution. These results are compared to those achieved by reference solutions (optimum and flooding-based), providing indications of the proposed solutions’ efficiency. These results are considered close to the optimum one and significantly better than the ones achieved by flooding-based solutions.

ITC/USA 2013 Conference Proceedings / The Forty-Ninth Annual International Telemetering Conference and Technical Exhibition / October 21-24, 2013 / Bally's Hotel & Convention Center, Las Vegas, NV / In this paper, a new technique utilizing a scatterer of high dielectric constant in between electrically small antennas to achieve good Direction of arrival (DOA) estimation performance is demonstrated. The phase information of the received signal at the antennas is utilized for direction estimation. The impact of the property of the scatterer on the directional sensitivity and the output signal to noise ratio (SNR) level are studied. Finally the DOA estimation accuracy is analyzed with the proposed technique under the consumption of white Gaussian noise environment.

Despite decades of effort in academia and industry, it is not yet possible to build machines that can replicate many seemingly-basic human perceptual abilities. This work focuses on the problem of face identification that most of us effortlessly solve daily. Substantial progress has been made towards the goal of automatically identifying faces under tightly controlled conditions; however, in the domain of unconstrained face images, many challenges remain. We observe that the recent combination of widespread digital photography, inexpensive digital storage and bandwidth, and online social networks has led to the sudden creation of repositories of billions of shared photographs and opened up an important new domain for unconstrained face identification research. Drawing upon the newly-popular phenomenon of “tagging,” we construct some of the first face identification datasets that are intended to model the digital social spheres of online social network members, and we examine various qualitative and quantitative properties of these image sets. The identification datasets we present here include up to 100 individuals, making them comparable to the average size of members’ networks of “friends” on a popular online social network, and each individual is represented by up to 100 face samples that feature significant real-world variation in appearance, expression, and pose. We demonstrate that biologically-inspired visual representations can achieve state-of-the-art face identification performance on our novel frontal and multi-pose face datasets. We also show that the addition of a tree-structured classifier and training set augmentation can enhance accuracy in the multi-pose setting. Finally, we illustrate that the machine-readable “social context” in which shared photos are often embedded can be applied to further boost face identification accuracy. Taken together, our results suggest that accurate automated face identification in vast online shared photo collections is now feasible. / Engineering and Applied Sciences

The main focus of this thesis is to develop biologically-based computational models for object recognition. A series of models for attention and object recognition were developed in the order of increasing functionality and complexity. These models are based on information processing in the primate brain, and specially inspired from the theory of visual information processing along the two parallel processing pathways of the primate visual cortex. To capture the true essence of incremental, constraint satisfaction style processing in the visual system, interactive neural networks were used for implementing our models. Results from eye-tracking studies on the relevant visual tasks, as well as our hypothesis regarding the information processing in the primate visual system, were implemented in the models and tested with simulations. As a first step, a model based on the ventral pathway was developed to recognize single objects. Through systematic testing, structural and algorithmic parameters of these models were fine tuned for performing their task optimally. In the second step, the model was extended by considering the dorsal pathway, which enables simulation of visual attention as an emergent phenomenon. The extended model was then investigated for visual search tasks. In the last step, we focussed on occluded and overlapped object recognition. A couple of eye-tracking studies were conducted in this regard and on the basis of the results we made some hypotheses regarding information processing in the primate visual system. The models were further advanced on the lines of the presented hypothesis, and simulated on the tasks of occluded and overlapped object recognition. On the basis of the results and analysis of our simulations we have further found that the generalization performance of interactive hierarchical networks improves with the addition of a small amount of Hebbian learning to an otherwise pure error-driven learning. We also concluded that the size of the receptive fields in our networks is an important parameter for the generalization task and depends on the object of interest in the image. Our results show that networks using hard coded feature extraction perform better than the networks that use Hebbian learning for developing feature detectors. We have successfully demonstrated the emergence of visual attention within an interactive network and also the role of context in the search task. Simulation results with occluded and overlapped objects support our extended interactive processing approach, which is a combination of the interactive and top-down approach, to the segmentation-recognition issue. Furthermore, the simulation behavior of our models is in line with known human behavior for similar tasks. In general, the work in this thesis will improve the understanding and performance of biologically-based interactive networks for object recognition and provide a biologically-plausible solution to recognition of occluded and overlapped objects. Moreover, our models provide some suggestions for the underlying neural mechanism and strategies behind biological object recognition.

Interactions are all around us, and as humans we may use words and gestures to communicate our intentions. At the micro level of fungi, communications are replaced by chemical signals and structure. These interactions fall into three distinctive categories: synergistic, where organisms help each other, as is the case with ectomycorrhizal fungi and tree roots, deadlock, or combat, where organisms fight for or defend a resource. When it comes to fungi-tree interactions, the fungi group of basidiomycetes fall into the latter category. At the onset of fungal infection, a living tree defends itself by producing resinous substances such as terpenes. These compounds are frequently found in hydrodistilled turpentine, which makes turpentine a prime source of antifungal compounds. A D-optimal design of fractionated turpentine together with gas chromatography (GC) coupled to a mass spectrometer was employed to find the most biologically active constituent of turpentine. Growth rate of Coniophora puteana was used to assess the efficacy of the mixed fractions. The partial least squares projection model had an excellent predictive power (R2 = 0.988, Q2 = 0.825) and validity. A putative sesquiterpene was identified as the most active compound for inhibiting fungal growth. The model was corroborated by an external validation assay employing preparative GC. After the death of a tree, fungi are no longer hindered by secondary metabolites from the tree. Instead, other interspecies interactions and intraspecies interactions, such as fungi-fungi interactions, occur. We found that when the white-rot fungus Heterobasidion parviporum and brown-rot fungus Gloeophyllum sepiarium interact with each other, amino acids are used to a higher extent. Amino acids may be used to produce antifungal compounds to hinder the other species from growing. Lysine in particular was utilized to a greater extent during interaction. Glutamine was the only amino acid that increased in concentration. Glutamine might be exuded or converted by enzymes from already existing glutamic acid. Dry weights suggest that the fungi were in a deadlock and that nutrient limitation might be a determining factor. It seemed that H. parviporum was favoured by a decrease in pH while the opposite pattern may be true for G. sepiarium.

Thesis (Ph.D)--Biomedical Engineering, Georgia Institute of Technology, 2008. / Committee Chair: Potter, Steve; Committee Member: Butera, Robert; Committee Member: DeMarse, Thomas; Committee Member: Jaeger, Dieter; Committee Member: Lee, Robert.

Thesis (Ph. D.)--Electrical and Computer Engineering, Georgia Institute of Technology, 2007. / Oliver Brand, Committee Member ; James H. McClellan, Committee Member ; Paul Hasler, Committee Chair ; Mark T. Smith, Committee Member ; David Anderson, Committee Member.

In this thesis, I describe a quantitative model that accounts for the circuits and computations of the feedforward path of the ventral stream of visual cortex. This model is consistent with a general theory of visual processing that extends the hierarchical model of (Hubel & Wiesel, 1959) from primary to extrastriate visual areas. It attempts to explain the first few hundred milliseconds of visual processing and “immediate recognition”. One of the key elements in the approach is the learning of a generic dictionary of shape-components from V2 to IT, which provides an invariant representation to task-specific categorization circuits in higher brain areas. This vocabulary of shape-tuned units is learned in an unsupervised manner from natural images, and constitutes a large and redundant set of image features with different complexities and invariances. This theory significantly extends an earlier approach by (Riesenhuber & Poggio, 1999) and builds upon several existing neurobiological models and conceptual proposals.First, I present evidence to show that the model can duplicate the tuning properties of neurons in various brain areas (e.g., V1, V4 and IT). In particular, the model agrees with data from V4 about the response of neurons to combinations of simple two-bar stimuli (Reynolds et al, 1999) (within the receptive field of the S2 units) and some of the C2 units in the model show a tuning for boundary conformations which is consistent with recordings from V4 (Pasupathy & Connor, 2001). Second, I show that not only can the model duplicate the tuning properties of neurons in various brain areas when probed with artificial stimuli, but it can also handle the recognition of objects in the real-world, to the extent of competing with the best computer vision systems. Third, I describe a comparison between the performance of the model and the performance of human observers in a rapid animal vs. non-animal recognition task for which recognition is fast and cortical back-projections are likely to be inactive. Results indicate that the model predicts human performance extremely well when the delay between the stimulus and the mask is about 50 ms. This suggests that cortical back-projections may not play a significant role when the time interval is in this range, and the model may therefore provide a satisfactory description of the feedforward path.Taken together, the evidences suggest that we may have the skeleton of a successful theory of visual cortex. In addition, this may be the first time that a neurobiological model, faithful to the physiology and the anatomy of visual cortex, not only competes with some of the best computer vision systems thus providing a realistic alternative to engineered artificial vision systems, but also achieves performance close to that of humans in a categorization task involving complex natural images. / PhD thesis