Um manãtodo eficiente de aprendizado de ma¡quina usa conhecimento quamico para criar uma grama¡tica que pode ser aprendida com regras de produção para construir mona´meros e polímeros sintetiza¡veis.
![](data:image/jpeg;base64,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)
Pesquisadores do MIT e da IBM usaram um modelo generativo com uma grama¡tica de grafos para criar novas moléculas pertencentes a mesma classe de compostos do conjunto de treinamento.
Engenheiros químicos e cientistas de materiais estãoconstantemente procurando o pra³ximo material revoluciona¡rio, produto quamico e medicamento. A ascensão das abordagens de aprendizado de ma¡quina estãoacelerando o processo de descoberta, que poderia levar anos. “Idealmente, o objetivo étreinar um modelo de aprendizado de ma¡quina em algumas amostras químicas existentes e permitir que ele produza o maior número possível de moléculas fabrica¡veis ​​da mesma classe, com propriedades físicas previsaveisâ€, diz Wojciech Matusik, professor de engenharia elanãtrica. e ciência da computação no MIT. “Se vocêtem todos esses componentes, pode construir novas moléculas com propriedades ideais e também sabe como sintetiza¡-las. Essa éa visão geral que as pessoas naquele espaço querem alcana§arâ€
No entanto, as técnicas atuais, principalmente o aprendizado profundo, exigem extensos conjuntos de dados para modelos de treinamento, e muitos conjuntos de dados químicos específicos de classe contem um punhado de compostos de exemplo, limitando sua capacidade de generalizar e gerar moléculas físicas que podem ser criadas no mundo real.
Agora, um novo artigo de pesquisadores do MIT e da IBM aborda esse problema usando um modelo de gra¡fico generativo para construir novas moléculas sintetiza¡veis ​​dentro da mesma classe química que seus dados de treinamento. Para fazer isso, eles tratam a formação de a¡tomos e ligações químicas como um gra¡fico e desenvolvem uma grama¡tica gra¡fica osuma analogia linguastica de sistemas e estruturas para ordenação de palavras osque contanãm uma sequaªncia de regras para construir molanãculas, como mona´meros e polímeros. Usando as regras gramaticais e de produção que foram inferidas do conjunto de treinamento, o modelo pode não apenas fazer engenharia reversa de seus exemplos, mas também criar novos compostos de maneira sistema¡tica e com eficiência de dados. “Basicamente construamos uma linguagem para criar molanãculasâ€, diz Matusik “Essa grama¡tica éessencialmente o modelo generativoâ€.
Os coautores de Matusik incluem os estudantes de pós-graduação do MIT Minghao Guo, que éo autor principal, e Beichen Li, bem como Veronika Thost, Payal Das e Jie Chen, membros da equipe de pesquisa da IBM Research. Matusik, Thost e Chen são afiliados ao MIT-IBM Watson AI Lab. Seu manãtodo, que eles chamaram de grama¡tica gra¡fica eficiente de dados (DEG), seráapresentado na Conferência Internacional sobre Representações de Aprendizagem.
“Queremos usar essa representação gramatical para geração de mona´meros e polímeros, porque essa grama¡tica éexplica¡vel e expressivaâ€, diz Guo. “Com apenas algumas regras de produção, podemos gerar muitos tipos de estruturas.â€
Uma estrutura molecular pode ser pensada como uma representação simba³lica em um gra¡fico osuma sequaªncia de a¡tomos (nós) unidos por ligações químicas (bordas). Nesse manãtodo, os pesquisadores permitem que o modelo pegue a estrutura química e colapse uma subestrutura da molanãcula atéum na³; podem ser dois a¡tomos ligados por uma ligação, uma sequaªncia curta de a¡tomos ligados ou um anel de a¡tomos. Isso éfeito repetidamente, criando as regras de produção a medida que avana§a, atéque um aºnico na³ permanea§a. As regras e grama¡tica poderiam então ser aplicadas na ordem inversa para recriar o conjunto de treinamento do zero ou combinados em diferentes combinações para produzir novas moléculas da mesma classe química.
“Os manãtodos de geração de grafos existentes produziriam um na³ ou uma aresta sequencialmente por vez, mas estamos analisando estruturas denívelsuperior e, especificamente, explorando o conhecimento de química, para que não tratemos os a¡tomos e ligações individuais como a unidade. Isso simplifica o processo de geração e também torna o aprendizado mais eficiente em dadosâ€, diz Chen.
Além disso, os pesquisadores otimizaram a técnica para que a grama¡tica de baixo para cima fosse relativamente simples e direta, de modo que fabricasse moléculas que pudessem ser feitas.
“Se mudarmos a ordem de aplicação dessas regras de produção, obteraamos outra molanãcula; além disso, podemos enumerar todas as possibilidades e gerar toneladas delasâ€, diz Chen. “Algumas dessas moléculas são va¡lidas e outras não, então o aprendizado da grama¡tica em si érealmente descobrir uma coleção manima de regras de produção, de modo que a porcentagem de moléculas que podem realmente ser sintetizadas seja maximizadaâ€. Enquanto os pesquisadores se concentraram em três conjuntos de treinamento com menos de 33 amostras cada osacrilatos, extensores de cadeia e isocianatos oseles observam que o processo pode ser aplicado a qualquer classe química.
Para ver como seu manãtodo funcionava, os pesquisadores testaram o DEG contra outros modelos e técnicas de última geração, analisando porcentagens de moléculas quimicamente va¡lidas e únicas, diversidade daquelas criadas, taxa de sucesso de retrossantese e porcentagem de moléculas pertencentes a a classe de mona´mero dos dados de treinamento.
“Mostramos claramente que, em termos de capacidade de santese e associação, nosso algoritmo supera todos os manãtodos existentes por uma margem muito grande, enquanto écompara¡vel a algumas outras manãtricas amplamente utilizadasâ€, diz Guo. Além disso, “o que éincravel em nosso algoritmo éque precisamos apenas de cerca de 0,15% do conjunto de dados original para obter resultados muito semelhantes em comparação com abordagens de última geração que treinam em dezenas de milhares de amostras. Nosso algoritmo pode lidar especificamente com o problema da escassez de dados.â€
No futuro imediato, a equipe planeja ampliar esse processo de aprendizado gramatical para poder gerar gra¡ficos grandes, bem como produzir e identificar produtos químicos com propriedades desejadas.
No futuro, os pesquisadores veem muitas aplicações para o manãtodo DEG, pois éadapta¡vel além de gerar novas estruturas químicas, aponta a equipe. Um gra¡fico éuma representação muito flexavel, e muitas entidades podem ser simbolizadas dessa forma osrobôs, veaculos, edifacios e circuitos eletra´nicos, por exemplo. “Essencialmente, nosso objetivo éconstruir nossa grama¡tica, para que nossa representação gra¡fica possa ser amplamente usada em muitos domanios diferentesâ€, diz Guo, pois “o DEG pode automatizar o design de novas entidades e estruturasâ€, diz Chen.
Esta pesquisa foi apoiada, em parte, pelo MIT-IBM Watson AI Lab e pela Evonik.