FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3147221285> ?p ?o ?g. }

• W3147221285 endingPage "2065" @default.
• W3147221285 startingPage "2065" @default.
• W3147221285 abstract "The question of molecular similarity is core in cheminformatics and is usually assessed via a pairwise comparison based on vectors of properties or molecular fingerprints. We recently exploited variational autoencoders to embed 6M molecules in a chemical space, such that their (Euclidean) distance within the latent space so formed could be assessed within the framework of the entire molecular set. However, the standard objective function used did not seek to manipulate the latent space so as to cluster the molecules based on any perceived similarity. Using a set of some 160,000 molecules of biological relevance, we here bring together three modern elements of deep learning to create a novel and disentangled latent space, viz transformers, contrastive learning, and an embedded autoencoder. The effective dimensionality of the latent space was varied such that clear separation of individual types of molecules could be observed within individual dimensions of the latent space. The capacity of the network was such that many dimensions were not populated at all. As before, we assessed the utility of the representation by comparing clozapine with its near neighbors, and we also did the same for various antibiotics related to flucloxacillin. Transformers, especially when as here coupled with contrastive learning, effectively provide one-shot learning and lead to a successful and disentangled representation of molecular latent spaces that at once uses the entire training set in their construction while allowing “similar” molecules to cluster together in an effective and interpretable way." @default.
• W3147221285 created "2021-04-13" @default.
• W3147221285 creator A5077370611 @default.
• W3147221285 creator A5081241139 @default.
• W3147221285 date "2021-04-03" @default.
• W3147221285 modified "2023-10-11" @default.
• W3147221285 title "FragNet, a Contrastive Learning-Based Transformer Model for Clustering, Interpreting, Visualizing, and Navigating Chemical Space" @default.
• W3147221285 cites W1534441087 @default.
• W3147221285 cites W1712092983 @default.
• W3147221285 cites W1757990252 @default.
• W3147221285 cites W1956858148 @default.
• W3147221285 cites W1975147762 @default.
• W3147221285 cites W1981802097 @default.
• W3147221285 cites W1990555250 @default.
• W3147221285 cites W1992156271 @default.
• W3147221285 cites W2000747708 @default.
• W3147221285 cites W2005066402 @default.
• W3147221285 cites W2008127220 @default.
• W3147221285 cites W2023818227 @default.
• W3147221285 cites W2034549041 @default.
• W3147221285 cites W2038952578 @default.
• W3147221285 cites W2047337557 @default.
• W3147221285 cites W2061741442 @default.
• W3147221285 cites W2076063813 @default.
• W3147221285 cites W2077280407 @default.
• W3147221285 cites W2077538176 @default.
• W3147221285 cites W2113362740 @default.
• W3147221285 cites W2115733720 @default.
• W3147221285 cites W2152826865 @default.
• W3147221285 cites W2153861051 @default.
• W3147221285 cites W2160592148 @default.
• W3147221285 cites W2168480393 @default.
• W3147221285 cites W2169863228 @default.
• W3147221285 cites W2174991771 @default.
• W3147221285 cites W2176290317 @default.
• W3147221285 cites W2176516200 @default.
• W3147221285 cites W2213443318 @default.
• W3147221285 cites W2316253021 @default.
• W3147221285 cites W2320034101 @default.
• W3147221285 cites W2412446857 @default.
• W3147221285 cites W2460166695 @default.
• W3147221285 cites W2506423325 @default.
• W3147221285 cites W2517981151 @default.
• W3147221285 cites W2549984318 @default.
• W3147221285 cites W2565684601 @default.
• W3147221285 cites W2594247694 @default.
• W3147221285 cites W2596398464 @default.
• W3147221285 cites W2738724892 @default.
• W3147221285 cites W2769484073 @default.
• W3147221285 cites W2770765812 @default.
• W3147221285 cites W2770849903 @default.
• W3147221285 cites W2791901977 @default.
• W3147221285 cites W2810417098 @default.
• W3147221285 cites W2883269759 @default.
• W3147221285 cites W2883583109 @default.
• W3147221285 cites W2889326414 @default.
• W3147221285 cites W2903782687 @default.
• W3147221285 cites W2906697496 @default.
• W3147221285 cites W2919115771 @default.
• W3147221285 cites W2920795827 @default.
• W3147221285 cites W2922332171 @default.
• W3147221285 cites W2943495267 @default.
• W3147221285 cites W2950838202 @default.
• W3147221285 cites W2958089299 @default.
• W3147221285 cites W2963518130 @default.
• W3147221285 cites W2966081953 @default.
• W3147221285 cites W2966284335 @default.
• W3147221285 cites W2966357564 @default.
• W3147221285 cites W2971690404 @default.
• W3147221285 cites W2973873632 @default.
• W3147221285 cites W2981731882 @default.
• W3147221285 cites W2995523160 @default.
• W3147221285 cites W2997728144 @default.
• W3147221285 cites W2999044305 @default.
• W3147221285 cites W3008695158 @default.
• W3147221285 cites W3011286504 @default.
• W3147221285 cites W3011609491 @default.
• W3147221285 cites W3023658436 @default.
• W3147221285 cites W3035302862 @default.
• W3147221285 cites W3045522182 @default.
• W3147221285 cites W3080764280 @default.
• W3147221285 cites W3083024963 @default.
• W3147221285 cites W3084673918 @default.
• W3147221285 cites W3092001068 @default.
• W3147221285 cites W3092931744 @default.
• W3147221285 cites W3093388624 @default.
• W3147221285 cites W3094771832 @default.
• W3147221285 cites W3094954720 @default.
• W3147221285 cites W3096655658 @default.
• W3147221285 cites W3098269892 @default.
• W3147221285 cites W3105920383 @default.
• W3147221285 cites W3107262343 @default.
• W3147221285 cites W3111772703 @default.
• W3147221285 cites W3111995631 @default.
• W3147221285 cites W3112559629 @default.
• W3147221285 cites W3113447514 @default.
• W3147221285 cites W3113596755 @default.
• W3147221285 cites W3117614253 @default.

• next