-
Analysis of Drug repurposing Knowledge graphs for Covid-19
Authors:
Ajay Kumar Gogineni
Abstract:
Knowledge graph (KG) is used to represent data in terms of entities and structural relations between the entities. This representation can be used to solve complex problems such as recommendation systems and question answering. In this study, a set of candidate drugs for COVID-19 are proposed by using Drug repurposing knowledge graph (DRKG). DRKG is a biological knowledge graph constructed using a…
▽ More
Knowledge graph (KG) is used to represent data in terms of entities and structural relations between the entities. This representation can be used to solve complex problems such as recommendation systems and question answering. In this study, a set of candidate drugs for COVID-19 are proposed by using Drug repurposing knowledge graph (DRKG). DRKG is a biological knowledge graph constructed using a vast amount of open source biomedical knowledge to understand the mechanism of compounds and the related biological functions. Node and relation embeddings are learned using knowledge graph embedding models and neural network and attention related models. Different models are used to get the node embedding by changing the objective of the model. These embeddings are later used to predict if a candidate drug is effective to treat a disease or how likely it is for a drug to bind to a protein associated to a disease which can be modelled as a link prediction task between two nodes. RESCAL performed the best on the test dataset in terms of MR, MRR and Hits@3.
△ Less
Submitted 7 December, 2022;
originally announced December 2022.
-
Multi-Class classification of vulnerabilities in Smart Contracts using AWD-LSTM, with pre-trained encoder inspired from natural language processing
Authors:
Ajay K. Gogineni,
S. Swayamjyoti,
Devadatta Sahoo,
Kisor K. Sahu,
Raj kishore
Abstract:
Vulnerability detection and safety of smart contracts are of paramount importance because of their immutable nature. Symbolic tools like OYENTE and MAIAN are typically used for vulnerability prediction in smart contracts. As these tools are computationally expensive, they are typically used to detect vulnerabilities until some predefined invocation depth. These tools require more search time as th…
▽ More
Vulnerability detection and safety of smart contracts are of paramount importance because of their immutable nature. Symbolic tools like OYENTE and MAIAN are typically used for vulnerability prediction in smart contracts. As these tools are computationally expensive, they are typically used to detect vulnerabilities until some predefined invocation depth. These tools require more search time as the invocation depth increases. Since the number of smart contracts is increasing exponentially, it is difficult to analyze the contracts using these traditional tools. Recently a machine learning technique called Long Short Term Memory (LSTM) has been used for binary classification, i.e., to predict whether a smart contract is vulnerable or not. This technique requires nearly constant search time as the invocation depth increases. In the present article, we have shown a multi-class classification, where we classify a smart contract in Suicidal, Prodigal, Greedy, or Normal categories. We used Average Stochastic Gradient Descent Weight-Dropped LSTM (AWD-LSTM), which is a variant of LSTM, to perform classification. We reduced the class imbalance (a large number of normal contracts as compared to other categories) by considering only the distinct opcode combination for normal contracts. We have achieved a weighted average Fbeta score of 90.0%. Hence, such techniques can be used to analyze a large number of smart contracts and help to improve the security of these contracts.
△ Less
Submitted 21 March, 2020;
originally announced April 2020.
-
Visual Machine Learning: Insight through Eigenvectors, Chladni patterns and community detection in 2D particulate structures
Authors:
Raj Kishore,
S. Swayamjyoti,
Shreeja Das,
Ajay K. Gogineni,
Zohar Nussinov,
D. Solenov,
Kisor K. Sahu
Abstract:
Machine learning (ML) is quickly emerging as a powerful tool with diverse applications across an extremely broad spectrum of disciplines and commercial endeavors. Typically, ML is used as a black box that provides little illuminating rationalization of its output. In the current work, we aim to better understand the generic intuition underlying unsupervised ML with a focus on physical systems. The…
▽ More
Machine learning (ML) is quickly emerging as a powerful tool with diverse applications across an extremely broad spectrum of disciplines and commercial endeavors. Typically, ML is used as a black box that provides little illuminating rationalization of its output. In the current work, we aim to better understand the generic intuition underlying unsupervised ML with a focus on physical systems. The systems that are studied here as test cases comprise of six different 2-dimensional (2-D) particulate systems of different complexities. It is noted that the findings of this study are generic to any unsupervised ML problem and are not restricted to materials systems alone. Three rudimentary unsupervised ML techniques are employed on the adjacency (connectivity) matrix of the six studied systems: (i) using principal eigenvalue and eigenvectors of the adjacency matrix, (ii) spectral decomposition, and (iii) a Potts model based community detection technique in which a modularity function is maximized. We demonstrate that, while solving a completely classical problem, ML technique produces features that are distinctly connected to quantum mechanical solutions. Dissecting these features help us to understand the deep connection between the classical non-linear world and the quantum mechanical linear world through the kaleidoscope of ML technique, which might have far reaching consequences both in the arena of physical sciences and ML.
△ Less
Submitted 2 January, 2020;
originally announced January 2020.
