Hierarchical Concept-Driven Language Model

For guiding natural language generation, many semantic-driven methods have been proposed. While clearly improving the performance of the end-to-end training task, these existing semantic-driven methods still have clear limitations: for example, (i) they only utilize shallow semantic signals (e.g., from topic models) with only a single stochastic hidden layer in their data generation process, which suffer easily from noise (especially adapted for short-text etc.) and lack of interpretation; (ii) they ignore the sentence order and document context, as they treat each document as a bag of sentences, and fail to capture the long-distance dependencies and global semantic meaning of a document. To overcome these problems, we propose a novel semantic-driven language modeling framework, which is a method to learn a Hierarchical Language Model and a Recurrent Conceptualization-enhanced Gamma Belief Network, simultaneously. For scalable inference, we develop the auto-encoding Variational Recurrent Inference, allowing efficient end-to-end training and simultaneously capturing global semantics from a text corpus. Especially, this article introduces concept information derived from high-quality lexical knowledge graph Probase, which leverages strong interpretability and anti-nose capability for the proposed model. Moreover, the proposed model captures not only intra-sentence word dependencies, but also temporal transitions between sentences and inter-sentence concept dependence. Experiments conducted on several NLP tasks validate the superiority of the proposed approach, which could effectively infer meaningful hierarchical concept structure of document and hierarchical multi-scale structures of sequences, even compared with latest state-of-the-art Transformer-based models.

Bayesian inference for compact binary coalescences with bilby: validation and application to the first LIGO–Virgo gravitational-wave transient catalogue

ABSTRACT Gravitational waves provide a unique tool for observational astronomy. While the first LIGO–Virgo catalogue of gravitational-wave transients (GWTC-1) contains 11 signals from black hole and neutron star binaries, the number of observations is increasing rapidly as detector sensitivity improves. To extract information from the observed signals, it is imperative to have fast, flexible, and scalable inference techniques. In a previous paper, we introduced bilby: a modular and user-friendly Bayesian inference library adapted to address the needs of gravitational-wave inference. In this work, we demonstrate that bilby produces reliable results for simulated gravitational-wave signals from compact binary mergers, and verify that it accurately reproduces results reported for the 11 GWTC-1 signals. Additionally, we provide configuration and output files for all analyses to allow for easy reproduction, modification, and future use. This work establishes that bilby is primed and ready to analyse the rapidly growing population of compact binary coalescence gravitational-wave signals.

PyClone-VI: Scalable inference of clonal population structures using whole genome data

We describe PyClone-VI, a computationally efficient Bayesian statistical method for inferring the clonal population structure of cancers. Our proposed method is 10-100x times faster than existing methods, while providing results which are as accurate. We demonstrate the utility of the method by analyzing data from 1717 patients from PCAWG study and 100 patients from the TRACERx study. Software implementing our method is freely available https://github.com/Roth-Lab/pyclone-vi.

CRISPRL and: Interpretable large-scale inference of DNA repair landscape based on a spectral approach

Summary We propose a new spectral framework for reliable training, scalable inference and interpretable explanation of the DNA repair outcome following a Cas9 cutting. Our framework, dubbed CRISPRL and, relies on an unexploited observation about the nature of the repair process: the landscape of the DNA repair is highly sparse in the (Walsh–Hadamard) spectral domain. This observation enables our framework to address key shortcomings that limit the interpretability and scaling of current deep-learning-based DNA repair models. In particular, CRISPRL and reduces the time to compute the full DNA repair landscape from a striking 5230 years to 1 week and the sampling complexity from 1012 to 3 million guide RNAs with only a small loss in accuracy (R2R2 ∼ 0.9). Our proposed framework is based on a divide-and-conquer strategy that uses a fast peeling algorithm to learn the DNA repair models. CRISPRL and captures lower-degree features around the cut site, which enrich for short insertions and deletions as well as higher-degree microhomology patterns that enrich for longer deletions. Availability and implementation The CRISPRL and software is publicly available at https://github.com/UCBASiCS/CRISPRLand.