Acculturation to Science: Becoming a Scientist in Mexico . The Challenge of Creating a Scientific Community in an Underdeveloped Country. Jacqueline Fortes and Larissa Adler Lomnitz. Pennsylvania State University Press, University Park, 1994. x, 225 pp., illus. $35 or £32. Translated from the Spanish by Alan P. Hynds.

Abstract

A key goal of the faculty was to transmit and inculcate into the students the model of an ideal scientist as combining hard work, discipline, and creativity. This -Ing a -. The Chdmodel was repeatedly emphasized in sem-) krnge of Creating a Scientii Community in an U m w JACQUUINE inars, in the laboratory, and by example K K~S ~L E R L O M N~. with the explicit aim of socializing stuv a n i a university w, University dents into a research culture. The program

MoTSE: an interpretable task similarity estimator for small molecular property prediction tasks

“…To further demonstrate the interpretability of MoTSE, we carried out a more challenging experiment, which included another five tasks of predicting the bioactivities of small molecules against cytochrome P450 isozymes. The cytochrome P450 (CYP) family plays important rules in drug metabolism, especially for five isozymes - 1A2, 2C9, 2C19, 2D6 and 3A4 [3, 37]. Here we obtained the binary bioactivity labels between around 17k molecules and the above five CYP isozymes from a preprocessed ChEMBL dataset [21,7,33].…”

“…RNN: RNN is a deep learning model particularly designed for processing sequential data, which has been proven to be effective in making molecular property predictions with one-hot encoded SMILES [1,2]. Here we employed a three-layer 128-dimensional LSTM [3] (a classical variant of RNN) to encode SMILES representations of molecules to 64-dimensional latent vectors and a two-layer (64-32-1) fully connected neural network to make prediction.…”