[postdoc] self-organizing hardware architectures
BISCUIT team, Loria
Supervision: Bernard Girau (HDR)
Applications will be considered until the position is filled.
N.B .: an Associate Professor position is likely to be published next year at Loria (for the start of the 2022 school year) with a research profile including neuromorphic approaches.
This post-doctorate proposal is in the context of a collaboration ranging from computational neurosciences to the design of reconfigurable multi-core circuits, and whose main goal is to study the possible contribution of the principles of neuronal structural plasticity within the framework of the design of massively distributed reconfigurable computational architectures.
As the quest for performance in computing systems confronts resource constraints, major breakthroughs in computing efficiency are expected to benefit from unconventional approaches and new models of computation such as brain-inspired computing. The brain is a massively parallel architecture with dense interconnections between computing units. Its structural organization is already a source of inspiration of several neuromorphic integrated circuits, but several dynamic properties of the brain computation have not really been explored from the point of view of a computational architecture. We are interested in the design of adaptive and dynamic computing architectures, taking advantage of brain-inspired self-organization principles. Self-organizing neural models have already been studied, especially for vector quantization tasks. However, existing self-organizing models are limited in the context of a self-organizing computing architecture project, because they rely on a predefined topology and they do not take into account communication costs nor afferent connectivity. Therefore, the transcription of self-organization at the hardware level is not straightforward and requires a number of challenges to be taken-up. The first challenge is to extend the usual self-organization mechanisms to account for the dual levels of computation and communication in a hardware neuromorphic architecture. From a biological point of view, this corresponds to a combination of the so-called synaptic and structural plasticities. Structural plasticity has been hardly considered in most self-organizing models because on one hand, such explicit dynamic lateral connectivity is computationally expensive and on the other hand, the exact rules governing the growing or the pruning of connections remain unclear.
Addressing these questions, on one hand we have defined a reconfigurable multi-core architecture (SCALP board, ) capable of exploiting the principles of hardware self-organization, and on the other hand we have defined different models of self-organizing maps integrating mechanisms of structural plasticity [2,3]. The objective of the work proposed here is to make the link between these two aspects:
– to study the implementation of self-organizing map models featuring structural plasticity on the SCALP board, by analyzing their behavior when dealing with real-world data and their properties with respect to the board’s communication constraints
– to interpret the vector quantization learned by the self-organizing maps in terms of communication needs among the computing units of the self-organizing architecture and in terms of dynamic allocation of computing resources within the SCALP board
The candidate must have the equivalent of a PhD in computer science, preferably on a subject related to artificial intelligence and / or distributed digital computing. Since most of the work will use a simulation software of the SCALP board, experienced skills in software engineering required, but knowledge of digital circuit design will also be taken into account. The candidate must be fluent in English and / or French.
 F. Vannel, D. Barrientos, J. Schmidt, C. Abegg, D. Buhlmann, andA. Upegui, “SCALP: Self-configurable 3-d cellular adaptive platform,”inProceedings of the 2018 IEEE Symposium Series on Computational Intelligence (SSCI), November 2018, pp. 1307–1312.
 B. Girau and A. Upegui, “Cellular self-organising maps,” inInt.Workshop on Self-Organizing Maps and Learning Vector Quantization,Clustering and Data Visualization (WSOM’19), 2019.
 A. Upegui, B. Girau, N. Rougier, F. Vannel, and B. Miramond, “Pruningself-organizing maps for cellular hardware architectures,” inNASA/ESAConf. on Adaptive Hardware and Systems (AHS 2018), August 2018.