检索结果-内蒙古大学图书馆

Adapting the deepSARM approach for dual-target ligand design

JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN 2021年第5期35卷 587-600页

作者： Yoshimori, Atsushi Hu, Huabin Bajorath, Jurgen Inst Theoret Med Inc 26-1 Muraoka Higashi 2 Chome Fujisawa Kanagawa 2510012 Japan Rhein Friedrich Wilhelms Univ LIMES Program Unit Chem Biol & Med Chem Dept Life Sci Informat B IT Friedrich Hirzebruch Allee 6 D-53115 Bonn Germany

The structure-activity relationship (SAR) matrix (SARM) methodology and data structure was originally developed to extract structurally related compound series from data sets of any composition, organize these series in matrices reminiscent of R-group tables, and visualize SAR patterns. The SARM approach combines the identification of structural relationships between series of active compounds with analog design, which is facilitated by systematically exploring combinations of core structures and substituents that have not been synthesized. The SARM methodology was extended through the introduction of deepSARM, which added deep learning and generative modeling to target-based analog design by taking compound information from related targets into account to further increase structural novelty. Herein, we present the foundations of the SARM methodology and discuss how deepSARM modeling can be adapted for the design of compounds with dual-target activity. Generating dual-target compounds represents an equally attractive and challenging task for polypharmacology-oriented drug discovery. The deepSARM-based approach is illustrated using a computational proof-of-concept application focusing on the design of candidate inhibitors for two prominent anti-cancer targets.

关键词： Structure– activity relationships SAR matrix Molecular grid maps deep generative modeling Dual-target compound design

来源：评论

学校读者我要写书评

暂无评论

generative AI for Sparse Antenna Array Design in ISAC Systems 25

Generative AI for Sparse Antenna Array Design in ISAC System...

引用

25th IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC)

作者： Evmorfos, Spilios Petropulu, Athina P. Rutgers State Univ Elect & Comp Engn New Brunswick NJ 08901 USA State Univ New Jersey Rutgers Elect & Comp Engn New Brunswick NJ USA

ISBN: (纸本)9798350393194;9798350393187

The fusion of sensing capabilities into communication systems stands as a pivotal advancement in wireless networking, giving rise to the paradigm of Integrated Sensing and Communication (ISAC) systems. In addition to the need for high performance in both sensing and communication functions, ISAC systems require energy efficiency and cost effectiveness. Antenna selection emerges as a strategic means to reduce power cost. The current work proposes a novel optimal antenna selection framework for ISAC systems, that is predicated on the deep generative modeling paradigm of generative Flow Networks (GFlowNets). Within this framework the antenna selection process is modeled as a deterministic Markov Decision Process (MDP) where sensor subsets arise as terminal states. Given that the objective comprises a convex combination of two terms, one for sensing and the other for communication, a multiobjective formulation is employed. This formulation conditions GFlowNet training on the vector of coefficients for both terms. The proposed approach not only optimizes the collective performance of the systems but also demonstrates generalization to scenarios with coefficients that are not used during

关键词： Antenna Selection ISAC GFlowNets deep generative modeling Multiobjective Optimization

来源：评论

学校读者我要写书评

暂无评论

LATENT REPRESENTATION AND CHARACTERIZATION OF SCANNING STRATEGY ON LASER POWDER BED FUSION ADDITIVE MANUFACTURING

LATENT REPRESENTATION AND CHARACTERIZATION OF SCANNING STRAT...

引用

ASME International Mechanical Engineering Congress and Exposition (IMECE)

作者： Imani, Farhad Chen, Ruimin Univ Connecticut Dept Mech Engn Storrs CT 06269 USA Univ Connecticut Dept Operat & Informat Management Storrs CT USA

ISBN: (纸本)9780791886649

Despite the transformative capability of laser powder bed fusion (LPBF) additive manufacturing to create components with intricate geometry, the large-scale adoption remains a barrier owing to the process complexity and significant build quality concerns. In-process melt pool imaging offers an unparalleled capability to tackle the problems by evaluating the impact of prominent process parameters (e.g., laser power, laser velocity, and hatch spacing) on build quality. However, the current investigations overlook the effect of other influential factors such as scan strategies. Because of the multitude and high-dimensionality in melt pool images, the extraction of manual features to characterize and intertwine diverse scan strategies (e.g., orthogonal serpentine, pre-scanned boarder, and clockwise spiral) is cumbersome or inefficient. While end-to-end deep neural networks realize automated feature extraction from melt pool images, they are limited in providing meaningful signatures for the characterization of various scan strategies. This paper presents a systematic image-guided analysis based on variational autoencoder (VAE) that enables the semantic representation of image data on low-dimensional latent space to characterize similarities between scan strategies. Further, hyperdimensional computing as a cognitive solution is integrated to differentiate various scan strategies according to latent features. Experimental results on the real-world case study based on 30,000 in-situ melt pool images show that VAE is significantly effective in interpretable characterization associated with 12 different scan strategies. In addition, the cognitive model differentiates scan strategies using the latent representation with an accuracy of 81.20 +/- 0.8%.

关键词： Laser Powder Bed Fusion Melt Pool Characterization Scan Strategies Analysis deep generative modeling

来源：评论

学校读者我要写书评

暂无评论

SpectraGAN: Spectrum based Generation of City Scale Spatiotemporal Mobile Network Traffic Data 21

SpectraGAN: Spectrum based Generation of City Scale Spatiote...

引用

17th ACM International Conference on Emerging Networking EXperiments and Technologies (ACM CoNEXT)

作者： Xu, Kai Singh, Rajkarn Fiore, Marco Marina, Mahesh K. Bilen, Hakan Usama, Muhammad Benn, Howard Ziemlicki, Cezary Univ Edinburgh Edinburgh Scotland IMDEA Networks Inst Madrid Spain Samsung Mountain View CA USA Orange Paris France

ISBN: (纸本)9781450390989

City-scale spatiotemporal mobile network traffic data can support numerous applications in and beyond networking. However, operators are very reluctant to share their data, which is curbing innovation and research reproducibility. To remedy this status quo, we propose SpectraGAN, a novel deep generative model that, upon training with real-world network traffic measurements, can produce high-fidelity synthetic mobile traffic data for new, arbitrary sized geographical regions over long periods. To this end, the model only requires publicly available context information about the target region, such as population census data. SpectraGAN is an original conditional GAN design with the defining feature of generating spectra of mobile traffic at all locations of the target region based on their contextual features. Evaluations with mobile traffic measurement datasets collected by different operators in 13 cities across two European countries demonstrate that SpectraGAN can synthesize more dependable traffic than a range of representative baselines from the literature. We also show that synthetic data generated with SpectraGAN yield similar results to that with real data when used in applications like radio access network infrastructure power savings and resource allocation, or dynamic population mapping.

关键词： Mobile Network Traffic Data Synthetic Data Generation deep generative modeling Conditional GANs

来源：评论

学校读者我要写书评

暂无评论

Two Souls in an Adversarial Image: Towards Universal Adversarial Example Detection using Multi-view Inconsistency 21

Two Souls in an Adversarial Image: Towards Universal Adversa...

引用

37th Annual Computer Security Applications Conference (ACSAC)

作者： Kiani, Sohaib Awan, Sana Lan, Chao Li, Fengjun Luo, Bo Univ Kansas Dept Elect Engn & Comp Sci Lawrence KS 66045 USA Univ Oklahoma Sch Comp Sci Norman OK 73019 USA

ISBN: (纸本)9781450385794

In the evasion attacks against deep neural networks (DNN), the attacker generates adversarial instances that are visually indistinguishable from benign samples and sends them to the target DNN to trigger misclassifications. In this paper, we propose a novel multi-view adversarial image detector, namely Argos, based on a novel observation. That is, there exist two "souls" in an adversarial instance, i.e., the visually unchanged content, which corresponds to the true label, and the added invisible perturbation, which corresponds to the misclassified label. Such inconsistencies could be further amplified through an autoregressive generative approach that generates images with seed pixels selected from the original image, a selected label, and pixel distributions learned from the training data. The generated images (i.e., the "views") will deviate significantly from the original one if the label is adversarial, demonstrating inconsistencies that Argos expects to detect. To this end, Argos first amplifies the discrepancies between the visual content of an image and its misclassified label induced by the attack using a set of regeneration mechanisms and then identifies an image as adversarial if the reproduced views deviate to a preset degree. Our experimental results show that Argos significantly outperforms two representative adversarial detectors in both detection accuracy and robustness against six well-known adversarial attacks. Code is available at: https://***/sohaib730/Argos-Adversarial_Detection

关键词： deep Learning Adversarial Attacks Adversarial Detection deep generative modeling Multi-view Machine Learning

来源：评论

学校读者我要写书评

暂无评论

GenDT: Mobile Network Drive Testing Made Efficient with generative modeling 22

GenDT: Mobile Network Drive Testing Made Efficient with Gene...

引用

18th International Conference on Emerging Networking Experiments and Technologies (ACM CoNEXT)

作者： Sun, Chuanhao Xu, Kai Marina, Mahesh K. Benn, Howard Univ Edinburgh Edinburgh Midlothian Scotland Samsung Suwon South Korea

ISBN: (纸本)9781450395083

Drive testing continues to play a key role in mobile network optimization for operators but its high cost is a big concern. Alternative approaches like virtual drive testing (VDT) target device testing in the lab whereas MDT or crowdsourcing based approaches are limited by the incentives users have to participate and contribute measurements. With the aim of augmenting drive testing and significantly reducing its cost, we propose GenDT, a novel deep generative model that synthesizes high-fidelity time series of key radio network key performance indicators (KPIs). The training of GenDT relies on a relatively small amount of real-world measurement data along with corresponding and easily accessible network and environment context data. Through this, GenDT learns the relationship between context and radio network KPIs as they vary over time, and therefore trained GenDT model can subsequently be relied on to generate time series for different KPIs for new drive test routes (trajectories) without having to collect field measurements. GenDT represents an initial attempt at enabling efficient drive testing via generative modeling. Evaluations with real-world mobile network drive testing measurement datasets from two countries demonstrate that GenDT can synthesize significantly more dependable data than a range of baselines. We further show that GenDT has the potential to significantly reduce the drive testing related measurement effort, and that GenDT-generated data yields similar results to that with real data in the context of two downstream use cases QoE prediction and handover analysis.

关键词： Mobile Network Drive Test Measurement Data Synthetic Data Generation deep generative modeling Conditional GANs

来源：评论

学校读者我要写书评

暂无评论

Graph2Plan: Learning Floorplan Generation from Layout Graphs

引用

ACM TRANSACTIONS ON GRAPHICS 2020年第4期39卷 118:1-118:14页

作者： Hu, Ruizhen Huang, Zeyu Tang, Yuhan van Kaick, Oliver Zhang, Hao Huang, Hui Shenzhen Univ Coll Comp Sci & Software Engn Shenzhen Guangdong Peoples R China Shenzhen Univ Shenzhen Guangdong Peoples R China Carleton Univ Ottawa ON Canada Simon Fraser Univ Burnaby BC Canada

We introduce a learning framework for automated floorplan generation which combines generative modeling using deep neural networks and userin-the-loop designs to enable human users to provide sparse design constraints. Such constraints are represented by a layout graph. The core component of our learning framework is a deep neural network, GRAPH2PLAN, which converts a layout graph, along with a building boundary, into a floor-plan that fulfills both the layout and boundary constraints. Given an input building boundary, we allow a user to specify room counts and other layout constraints, which are used to retrieve a set of floorplans, with their associated layout graphs, from a database. For each retrieved layout graph, along with the input boundary, GRAPH2PLAN first generates a corresponding raster floorplan image, and then a refined set of boxes representing the rooms. GRAPH2PLAN is trained on RPLAN, a large-scale dataset consisting of 80K annotated floorplans. The network is mainly based on convolutional processing over both the layout graph, via a graph neural network (GNN), and the input building boundary, as well as the raster floorplan images, via conventional image convolution. We demonstrate the quality and versatility of our floorplan generation framework in terms of its ability to cater to different user inputs. We conduct both qualitative and quantitative evaluations, ablation studies, and comparisons with state-of-the-art approaches.

关键词： floorplan generation layout graph deep generative modeling

来源：评论

学校读者我要写书评

暂无评论

Design subspace learning: Structural design space exploration using performance-conditioned generative modeling

引用

AUTOMATION IN CONSTRUCTION 2021年 127卷

作者： Danhaive, Renaud Mueller, Caitlin T. MIT Bldg Technol Program Dept Architecture 77 Massachusetts Ave Cambridge MA 02139 USA

Designers increasingly rely on parametric design studies to explore and improve structural concepts based on quantifiable metrics, generally either by generating design variations manually or using optimization methods. Unfortunately, both of these approaches have important shortcomings: effectively searching a large design space manually is infeasible, and design optimization overlooks qualitative aspects important in architectural and structural design. There is a need for methods that take advantage of computing intelligence to augment a designer's creativity while guiding-not forcing-their search for better-performing solutions. This research addresses this need by integrating conditional variational autoencoders in a performance-driven design exploration framework. First, a sampling algorithm generates a dataset of meaningful design options from an unwieldy design space. Second, a performance-conditioned variational autoencoder with a low-dimensional latent space is trained using the collected data. This latent space is intuitive to explore by designers even as it offers a diversity of high-performing design options.

关键词： deep generative modeling Latent space Latent variable Variational autoencoder Design space Computational design

来源：评论

学校读者我要写书评

暂无评论

A multimodal dynamical variational autoencoder for audiovisual speech representation learning

引用

NEURAL NETWORKS 2024年 172卷 106120页

作者： Sadok, Samir Leglaive, Simon Girin, Laurent Alameda-Pineda, Xavier Seguier, Renaud CentraleSupelec UMR CNRS 6164 IETR Gif Sur Yvette France Univ Grenoble Alpes CNRS Grenoble INP GIPSA Lab Grenoble France Univ Grenoble Alpes Inria LJK CNRS Grenoble France

High -dimensional data such as natural images or speech signals exhibit some form of regularity, preventing their dimensions from varying independently. This suggests that there exists a lower dimensional latent representation from which the high -dimensional observed data were generated. Uncovering the hidden explanatory features of complex data is the goal of representation learning, and deep latent variable generative models have emerged as promising unsupervised approaches. In particular, the variational autoencoder (VAE) which is equipped with both a generative and an inference model allows for the analysis, transformation, and generation of various types of data. Over the past few years, the VAE has been extended to deal with data that are either multimodal or dynamical (i.e., sequential). In this paper, we present a multimodal and dynamical VAE (MDVAE) applied to unsupervised audiovisual speech representation learning. The latent space is structured to dissociate the latent dynamical factors that are shared between the modalities from those that are specific to each modality. A static latent variable is also introduced to encode the information that is constant over time within an audiovisual speech sequence. The model is trained in an unsupervised manner on an audiovisual emotional speech dataset, in two stages. In the first stage, a vector quantized VAE (VQ-VAE) is learned independently for each modality, without temporal modeling. The second stage consists in learning the MDVAE model on the intermediate representation of the VQ-VAEs before quantization. The disentanglement between static versus dynamical and modality -specific versus modality -common information occurs during this second training stage. Extensive experiments are conducted to investigate how audiovisual speech latent factors are encoded in the latent space of MDVAE. These experiments include manipulating audiovisual speech, audiovisual facial image denoising, and audiovisual speech emotion

关键词： deep generative modeling Disentangled representation learning Variational autoencoder Multimodal and dynamical data Audiovisual speech processing

来源：评论

学校读者我要写书评

暂无评论

Vitruvio: Conditional variational autoencoder to generate building meshes via single perspective sketches

引用

AUTOMATION IN CONSTRUCTION 2024年 166卷

作者： Tono, Alberto Huang, Heyaojing Agrawal, Ashwin Fischer, Martin Stanford Univ Palo Alto CA USA Computat Design Inst San Francisco CA USA

At the beginning of a project, architects convey design ideas via quick 2D diagrams, front views, floor plans, and sketches. Consequently, many stakeholders have difficulty visualizing the 3D representation of the building mass, leading to varied interpretations thus inhibiting a shared understanding of the design. To alleviate the challenge, this paper proposes a deep learning-based method, Vitruvio, for creating a 3D model from a single perspective sketch. This method allows designers to automatically generate 3D representations in real-time based on their initial sketches and thus communicate effectively and intuitively to the client. Vitruvio adapts the Occupancy Network to perform single view reconstruction (SVR), a technique for creating 3D representations from a single image. Vitruvio achieves: (1) an 18% increase in the reconstruction accuracy and (2) a 26% reduction in the inference time compared to the Occupancy Network on one thousand buildings provided by the New York municipality. This research investigates the effect that the building orientation has on the reconstruction quality, discovering that Vitruvio can capture fine-grain details in complex buildings when their native orientation is preserved during training, as opposed to the SVR's standard practice that aligns every building to its canonical pose. The code is available here https://***/CDInstitute/Vitruvio.

关键词： Artificial intelligence Neural -aided design deep generative design deep generative modeling Conditional variational autoencoder Sketch -based modeling

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：