More Like this

1. End-to-end sequence-structure-function meta-learning predicts genome-wide chemical-protein interactions for dark proteins

   https://doi.org/10.1371/journal.pcbi.1010851  

   Cai, Tian ; Xie, Li ; Zhang, Shuo ; Chen, Muge ; He, Di ; Badkul, Amitesh ; Liu, Yang ; Namballa, Hari Krishna ; Dorogan, Michael ; Harding, Wayne W. ; et al ( January 2023 , PLOS Computational Biology)

   Skolnick, Jeffrey (Ed.)

   Systematically discovering protein-ligand interactions across the entire human and pathogen genomes is critical in chemical genomics, protein function prediction, drug discovery, and many other areas. However, more than 90% of gene families remain “dark”—i.e., their small-molecule ligands are undiscovered due to experimental limitations or human/historical biases. Existing computational approaches typically fail when the dark protein differs from those with known ligands. To address this challenge, we have developed a deep learning framework, called PortalCG, which consists of four novel components: (i) a 3-dimensional ligand binding site enhanced sequence pre-training strategy to encode the evolutionary links between ligand-binding sites across gene families; (ii) an end-to-end pretraining-fine-tuning strategy to reduce the impact of inaccuracy of predicted structures on function predictions by recognizing the sequence-structure-function paradigm; (iii) a new out-of-cluster meta-learning algorithm that extracts and accumulates information learned from predicting ligands of distinct gene families (meta-data) and applies the meta-data to a dark gene family; and (iv) a stress model selection step, using different gene families in the test data from those in the training and development data sets to facilitate model deployment in a real-world scenario. In extensive and rigorous benchmark experiments, PortalCG considerably outperformed state-of-the-art techniques of machine learning and protein-ligand docking when applied to dark gene families, and demonstrated its generalization power for target identifications and compound screenings under out-of-distribution (OOD) scenarios. Furthermore, in an external validation for the multi-target compound screening, the performance of PortalCG surpassed the rational design from medicinal chemists. Our results also suggest that a differentiable sequence-structure-function deep learning framework, where protein structural information serves as an intermediate layer, could be superior to conventional methodology where predicted protein structures were used for the compound screening. We applied PortalCG to two case studies to exemplify its potential in drug discovery: designing selective dual-antagonists of dopamine receptors for the treatment of opioid use disorder (OUD), and illuminating the understudied human genome for target diseases that do not yet have effective and safe therapeutics. Our results suggested that PortalCG is a viable solution to the OOD problem in exploring understudied regions of protein functional space. 

   more » « less
2. Proof transfer for fast certification of multiple approximate neural networks

   https://doi.org/10.1145/3527319  

   Ugare, Shubham ; Singh, Gagandeep ; Misailovic, Sasa ( April 2022 , Proceedings of the ACM on Programming Languages)

   Developers of machine learning applications often apply post-training neural network optimizations, such as quantization and pruning, that approximate a neural network to speed up inference and reduce energy consumption, while maintaining high accuracy and robustness. Despite a recent surge in techniques for the robustness verification of neural networks, a major limitation of almost all state-of-the-art approaches is that the verification needs to be run from scratch every time the network is even slightly modified. Running precise end-to-end verification from scratch for every new network is expensive and impractical in many scenarios that use or compare multiple approximate network versions, and the robustness of all the networks needs to be verified efficiently. We present FANC, the first general technique for transferring proofs between a given network and its multiple approximate versions without compromising verifier precision. To reuse the proofs obtained when verifying the original network, FANC generates a set of templates – connected symbolic shapes at intermediate layers of the original network – that capture the proof of the property to be verified. We present novel algorithms for generating and transforming templates that generalize to a broad range of approximate networks and reduce the verification cost. We present a comprehensive evaluation demonstrating the effectiveness of our approach. We consider a diverse set of networks obtained by applying popular approximation techniques such as quantization and pruning on fully-connected and convolutional architectures and verify their robustness against different adversarial attacks such as adversarial patches, L 0 , rotation and brightening. Our results indicate that FANC can significantly speed up verification with state-of-the-art verifier, DeepZ by up to 4.1x. 

   more » « less
3. Improving second language vocabulary learning and retention by leveraging memory enhancement techniques: A multidomain pedagogical approach

   https://doi.org/10.1177/13621688211053525  

   Karatas, Nur Basak ; Özemir, Oya ; Lovelett, Jarrett T. ; Demir, Bora ; Erkol, Kemal ; Veríssimo, João ; Erçetin, Gülcan ; Ullman, Michael T. ( October 2021 , Language Teaching Research)

   We investigated whether learning and retaining vocabulary in a second language (L2) can be improved by leveraging a combination of memory enhancement techniques. Specifically, we tested whether combining retrieval practice, spacing, and related manipulations in a ‘multidomain’ pedagogical approach enhances vocabulary acquisition as compared to a typical learning approach. In a classroom-laboratory design, 48 Turkish university students studying L2 English were trained on 64 English words over 17 days. They were assigned to either a ‘typical’ study regimen of (re)studying the words on the first day (initial study) and last day (cramming) of training, or an ‘optimized’ regimen of retrieval practice (retrieving the words), moreover with feedback, spaced throughout the period, moreover with expanding gaps. The target words were tested before training (pre-test) and one and 11 days afterwards (post-tests). Mixed-effects modeling revealed a training-group by test-session interaction, due to greater improvements from optimized training (a striking 18 percentage-point accuracy increase from pre-test to both post-tests) than typical training (an 8 percentage-point increase). Further analyses showed that the optimized training advantages were mainly driven by high (rather than low) frequency words. Overall, the results suggest that a multidomain approach of combining different memory enhancement techniques can lead to substantial gains in both the learning and retention of L2 words, as compared to a typical study pattern. The findings have implications for L2 learning and pedagogy.

    

   more » « less
4. Traceability Support for Multi-Lingual Software Projects

   https://doi.org/10.1145/3379597.3387440  

   Liu, Yalin ; Lin, Jinfeng ; Cleland-Huang, Jane ( July 2020 , International Conference on Mining Software Repositories)

   Software traceability establishes associations between diverse software artifacts such as requirements, design, code, and test cases. Due to the non-trivial costs of manually creating and maintaining links, many researchers have proposed automated approaches based on information retrieval techniques. However, many globally distributed software projects produce software artifacts written in two or more languages. The use of intermingled languages reduces the efficacy of automated tracing solutions. In this paper, we first analyze and discuss patterns of intermingled language use across multiple projects, and then evaluate several different tracing algorithms including the Vector Space Model (VSM), Latent Semantic Indexing (LSI), Latent Dirichlet Allocation (LDA), and various models that combine mono- and cross-lingual word embeddings with the Generative Vector Space Model (GVSM). Based on an analysis of 14 Chinese-English projects, our results show that best performance is achieved using mono-lingual word embeddings integrated into GVSM with machine translation as a preprocessing step. 

   more » « less
5. Traceability Transformed: Generating More Accurate Links with Pre-Trained BERT Models

   https://doi.org/10.1109/ICSE43902.2021.00040  

   Lin, Jinfeng ; Liu, Yalin ; Zeng, Qingkai ; Jiang, Meng ; Cleland-Huang, Jane ( May 2021 , Proceedings of the 43rd International Conference on Sofware Engineering)

   null (Ed.)

   Software traceability establishes and leverages associations between diverse development artifacts. Researchers have proposed the use of deep learning trace models to link natural language artifacts, such as requirements and issue descriptions, to source code; however, their effectiveness has been restricted by availability of labeled data and efficiency at runtime. In this study, we propose a novel framework called Trace BERT (T-BERT) to generate trace links between source code and natural language artifacts. To address data sparsity, we leverage a three-step training strategy to enable trace models to transfer knowledge from a closely related Software Engineering challenge, which has a rich dataset, to produce trace links with much higher accuracy than has previously been achieved. We then apply the T-BERT framework to recover links between issues and commits in Open Source Projects. We comparatively evaluated accuracy and efficiency of three BERT architectures. Results show that a Single-BERT architecture generated the most accurate links, while a Siamese-BERT architecture produced comparable results with significantly less execution time. Furthermore, by learning and transferring knowledge, all three models in the framework outperform classical IR trace models. On the three evaluated real-word OSS projects, the best T-BERT stably outperformed the VSM model with average improvements of 60.31% measured using Mean Average Precision (MAP). RNN severely underperformed on these projects due to insufficient training data, while T-BERT overcame this problem by using pretrained language models and transfer learning. 

   more » « less