Search

Proteins adopt multiple structural conformations to perform their diverse biological functions, and understanding these conformations is crucial for advancing drug discovery. Traditional physics-based simulation methods often struggle with sampling equilibrium conformations and are computationally expensive. Recently, deep generative models have shown promise in generating protein conformations as a more efficient alternative. However, these methods predominantly rely on the diffusion process within a 3D geometric space, which typically centers around the vicinity of metastable states and is often inefficient in terms of runtime. In this paper, we introduce Structure Language Modeling (SLM) as a novel framework for efficient protein conformation generation. Specifically, the protein structures are first encoded into a compact latent space using a discrete variational auto-encoder, followed by conditional language modeling that effectively captures sequence-specific conformation distributions. This enables a more efficient and interpretable exploration of diverse ensemble modes compared to existing methods. Based on this general framework, we instantiate SLM with various popular LM architectures as well as proposing the ESMDiff, a novel BERT-like structure language model fine-tuned from ESM3 with masked diffusion. We verify our approach in various scenarios, including the equilibrium dynamics of BPTI, conformational change pairs, and intrinsically disordered proteins. SLM provides a highly efficient solution, offering a 20-100x speedup than existing methods in generating diverse conformations, shedding light on promising avenues for future research., Comment: Preprint. Under Review

High-content phenotypic screening, including high-content imaging (HCI), has gained popularity in the last few years for its ability to characterize novel therapeutics without prior knowledge of the protein target. When combined with deep learning techniques to predict and represent molecular-phenotype interactions, these advancements hold the potential to significantly accelerate and enhance drug discovery applications. This work focuses on the novel task of HCI-guided molecular design. Generative models for molecule design could be guided by HCI data, for example with a supervised model that links molecules to phenotypes of interest as a reward function. However, limited labeled data, combined with the high-dimensional readouts, can make training these methods challenging and impractical. We consider an alternative approach in which we leverage an unsupervised multimodal joint embedding to define a latent similarity as a reward for GFlowNets. The proposed model learns to generate new molecules that could produce phenotypic effects similar to those of the given image target, without relying on pre-annotated phenotypic labels. We demonstrate that the proposed method generates molecules with high morphological and structural similarity to the target, increasing the likelihood of similar biological activity, as confirmed by an independent oracle model.

The growing popularity of generative AI, particularly ChatGPT, has sparked both enthusiasm and caution among practitioners and researchers in education. To effectively harness the full potential of ChatGPT in educational contexts, it is crucial to analyze its impact and suitability for different educational purposes. This paper takes an initial step in exploring the applicability of ChatGPT in a computer-supported collaborative learning (CSCL) environment. Using statistical analysis, we validate the shifts in student interactions during an asynchronous group brainstorming session by introducing ChatGPT as an instantaneous question-answering agent., Comment: AAAI2024 Workshop on AI for Education (AI4ED)

Generative Flow Networks (GFlowNets; GFNs) are a family of energy-based generative methods for combinatorial objects, capable of generating diverse and high-utility samples. However, consistently biasing GFNs towards producing high-utility samples is non-trivial. In this work, we leverage connections between GFNs and reinforcement learning (RL) and propose to combine the GFN policy with an action-value estimate, $Q$, to create greedier sampling policies which can be controlled by a mixing parameter. We show that several variants of the proposed method, QGFN, are able to improve on the number of high-reward samples generated in a variety of tasks without sacrificing diversity., Comment: Accepted by 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

Due to the steady rise in population demographics and longevity, emergency department visits are increasing across North America. As more patients visit the emergency department, traditional clinical workflows become overloaded and inefficient, leading to prolonged wait-times and reduced healthcare quality. One of such workflows is the triage medical directive, impeded by limited human workload, inaccurate diagnoses and invasive over-testing. To address this issue, we propose TriNet: a machine learning model for medical directives that automates first-line screening at triage for conditions requiring downstream testing for diagnosis confirmation. To verify screening potential, TriNet was trained on hospital triage data and achieved high positive predictive values in detecting pneumonia (0.86) and urinary tract infection (0.93). These models outperform current clinical benchmarks, indicating that machine-learning medical directives can offer cost-free, non-invasive screening with high specificity for common conditions, reducing the risk of over-testing while increasing emergency department efficiency., Comment: Index Terms: Downstream testing, Machine Learning, Medical directives, Modelling, Modular network, Pneumonia, Positive predictive value, Screening, Triage, Urinary tract infection

Matrix completion has a long-time history of usage as the core technique of recommender systems. In particular, 1-bit matrix completion, which considers the prediction as a ``Recommended'' or ``Not Recommended'' question, has proved its significance and validity in the field. However, while customers and products aggregate into interacted clusters, state-of-the-art model-based 1-bit recommender systems do not take the consideration of grouping bias. To tackle the gap, this paper introduced Group-Specific 1-bit Matrix Completion (GS1MC) by first-time consolidating group-specific effects into 1-bit recommender systems under the low-rank latent variable framework. Additionally, to empower GS1MC even when grouping information is unobtainable, Cluster Developing Matrix Completion (CDMC) was proposed by integrating the sparse subspace clustering technique into GS1MC. Namely, CDMC allows clustering users/items and to leverage their group effects into matrix completion at the same time. Experiments on synthetic and real-world data show that GS1MC outperforms the current 1-bit matrix completion methods. Meanwhile, it is compelling that CDMC can successfully capture items' genre features only based on sparse binary user-item interactive data. Notably, GS1MC provides a new insight to incorporate and evaluate the efficacy of clustering methods while CDMC can be served as a new tool to explore unrevealed social behavior or market phenomenon., Comment: 16 Pages

The agile approach uses continuous delivery, instead of distinct procedure, to work closer with customers and to respond faster requirement changes. All of these are against the traditional plan driven approach. Due to agile method's characteristics and its success in the real world practices, a number of discussions regarding the differences between agile and traditional approaches emerged recently and many studies intended to integrate both methods to synthesize the benefits from these two sides. However, this type of research often concludes from observations of a development activity or surveys after a project. To provide a more objective supportive evidence of comparing these two approaches, our research analyzes the source codes, logs, and notes. We argue that the agile and traditional approaches share common characteristics, which can be considered as the glue for integrating both methods. In our study, we collect all the submissions from the version control repository, and meeting notes and discussions. By applying our suggested analysis method, we illustrate the shared properties between agile and traditional approaches; thus, different development phases, like implementation and test, can still be identified in agile development history. This result not only provides a positive result for our hypothesis but also offers a suggestion for a better integration., Comment: 6 pages

An n-dimensional complex manifold M is said to be (holomorphically) dominable by $\CC^n$ if there is a map $F:\CC^n \ra M$ which is holomorphic such that the Jacobian determinant $\det(DF)$ is not identically zero. Such a map F is called a dominating map. In this paper, we attempt to classify algebraic surfaces X which are dominable by $\CC^2$ using a combination of techniques from algebraic topology, complex geometry and analysis. One of the key tools in the study of algebraic surfaces is the notion of Kodaira dimension (defined in section 2). By Kodaira's pioneering work and its extensions, an algebraic surface which is dominable by $\CC^2$ must have Kodaira dimension less than two. Using the Kodaira dimension and the fundamental group of X, we succeed in classifying algebraic surfaces which are dominable by $\CC^2$ except for certain cases in which X is an algebraic surface of Kodaira dimension zero and the case when X is rational without any logarithmic 1-form. More specifically, in the case when X is compact (namely projective), we need to exclude only the case when X is birationally equivalent to a K3 surface (a simply connected compact complex surface which admits a globally non-vanishing holomorphic 2-form) that is neither elliptic nor Kummer (see sections 3 and 4 for the definition of these types of surfaces)., Comment: 39 pages

Using the Kodaira dimension and the fundamental group of X, we succeed in classifying algebraic surfaces which are dominable by C^2 except for certain cases in which X is an algebraic surface of Kodaira dimension zero and the case when X is rational without any logarithmic 1-form. More specifically, in the case when X is compact (namely projective), we need to exclude only the case when X is birationally equivalent to a K3 surface (a simply connected compact complex surface which admits a globally non-vanishing holomorphic 2-form) that is neither elliptic nor Kummer. With the exceptions noted above, we show that for any algebraic surface of Kodaira dimension less than 2, dominability by C^2 is equivalent to the apparently weaker requirement of the existence of a holomorphic image of C which is Zariski dense in the surface. With the same exceptions, we will also show the very interesting and revealing fact that dominability by C^2 is preserved even if a sufficiently small neighborhood of any finite set of points is removed from the surface. In fact, we will provide a complete classification in the more general category of (not necessarily algebraic) compact complex surfaces before tackling the problem in the case of non-compact algebraic surfaces.