Search

Digital circuits are one of the most important enabling technologies in the world today. Powerful tools, such as Hardware Description Languages (HDLs) have evolved over the past number of decades to allow designers to operate at high levels of abstraction and expressiveness, rather than at the gate level, which circuits are actually constructed from. Similarly, highly accurate digital circuit simulators permit designers to test their circuits before committing them to silicon. This is still a highly complex and generally manual task, however, with complex circuits taking months or even years to go from planning to silicon. We show how Grammatical Evolution (GE) can harness the standard tools of silicon design and be used to create a fully automatic circuit design system. Specifically, we use a HDL known as SystemVerilog and Icarus, a free, but powerful simulator, to generate circuits from high level descriptions. We apply our system to several well known digital circuit literature benchmarks and demonstrate that GE can successfully evolve functional circuits, including several which have been subsequently rendered in Field Programmable Gate Arrays (FPGAs).

The association of the anti-tuberculosis drug rifampicin (RIF) with a 4th-generation poly(amidoamine) (G4-PAMAM) dendrimer was investigated by means of molecular dynamics simulations. The RIF load capacity was estimated to be around 20 RIF per G4-PAMAM at neutral pH. The complex formed by 20 RIF molecules and the dendrimer (RIF20-PAMAM) was subjected to 100 ns molecular dynamics (MD) simulations at two different pH conditions (neutral and acidic). The complex was found to be significantly more stable in the simulation at neutral pH compared to the simulation at low pH in which the RIF molecules were rapidly and almost simultaneously expelled to the solvent bulk. The high stability of the RIF-PAMAM complex under physiological pH and the rapid release of RIF molecules under acidic medium provide an interesting switch for drug targeting since the Mycobacterium resides within acidic domains of the macrophage. Altogether, these results suggest that, at least in terms of stability and pH-dependent release, PAMAM-like dendrimers may be considered suitable drug delivery systems for RIF and derivatives.

PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO As redes neurais deep são modelos poderosos e flexíveis, que ganharam destaque na comunidade científica na última década. Para muitas tarefas, elas até superam o desempenho humano. Em geral, para obter tais resultados, um especialista despende tempo significativo para projetar a arquitetura neural, com longas sessões de tentativa e erro. Com isso, há um interesse crescente em automatizar esse processo. Novos métodos baseados em técnicas como aprendizado por reforço e algoritmos evolutivos foram apresentados como abordagens para o problema da busca de arquitetura neural (NAS - Neural Architecture Search), mas muitos ainda são algoritmos de alto custo computacional. Para reduzir esse custo, pesquisadores sugeriram limitar o espaço de busca, com base em conhecimento prévio. Os algoritmos evolutivos de inspiração quântica (AEIQ) apresentam resultados promissores em relação à convergência mais rápida. A partir dessa idéia, propõe-se o Q-NAS: um AEIQ para buscar redes deep através da montagem de subestruturas. O Q-NAS também pode evoluir alguns hiperparâmetros numéricos, o que é um primeiro passo para a automação completa. Experimentos com o conjunto de dados CIFAR-10 foram realizados a fim de analisar detalhes do Q-NAS. Para muitas configurações de parâmetros, foram obtidos resultados satisfatórios. As melhores acurácias no CIFAR-10 foram de 93,85 porcento para uma rede residual e 93,70 porcento para uma rede convolucional, superando modelos elaborados por especialistas e alguns métodos de NAS. Incluindo um esquema simples de parada antecipada, os tempos de evolução nesses casos foram de 67 dias de GPU e 48 dias de GPU, respectivamente. O Q-NAS foi aplicado ao CIFAR-100, sem qualquer ajuste de parâmetro, e obteve 74,23 porcento de acurácia, similar a uma ResNet com 164 camadas. Por fim, apresenta-se um estudo de caso com dados reais, no qual utiliza-se o Q-NAS para resolver a tarefa de classificação sísmica. Em menos de 8,5 dias de GPU, o Q-NAS gerou redes com 12 vezes menos pesos e maior acurácia do que um modelo criado especialmente para esta tarefa. Deep neural networks are powerful and flexible models that have gained the attention of the machine learning community over the last decade. For a variety of tasks, they can even surpass human-level performance. Usually, to reach these excellent results, an expert spends significant time designing the neural architecture, with long trial and error sessions. In this scenario, there is a growing interest in automating this design process. To address the neural architecture search (NAS) problem, authors have presented new methods based on techniques such as reinforcement learning and evolutionary algorithms, but the high computational cost is still an issue for many of them. To reduce this cost, researchers have proposed to restrict the search space, with the help of expert knowledge. Quantum-inspired evolutionary algorithms present promising results regarding faster convergence. Motivated by this idea, we propose Q-NAS: a quantum-inspired algorithm to search for deep networks by assembling substructures. Q-NAS can also evolve some numerical hyperparameters, which is a first step in the direction of complete automation. We ran several experiments with the CIFAR-10 dataset to analyze the details of the algorithm. For many parameter settings, Q-NAS was able to achieve satisfactory results. Our best accuracies on the CIFAR-10 task were 93.85 percent for a residual network and 93.70 percent for a convolutional network, overcoming hand-designed models, and some NAS works. Considering the addition of a simple early-stopping mechanism, the evolution times for these runs were 67 GPU days and 48 GPU days, respectively. Also, we applied Q-NAS to CIFAR-100 without any parameter adjustment, reaching an accuracy of 74.23 percent, which is comparable to a ResNet with 164 layers. Finally, we present a case study with real datasets, where we used Q-NAS to solve the seismic classification task. In less than 8.5 GPU days, Q-NAS generated networks with 12 times fewer weights and higher accuracy than a model specially created for this task.