检索结果-内蒙古大学图书馆

Attention-based neural load forecasting: A dynamic feature selection approach

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Xiong, Jing Zhou, Pengyang Chen, Alan Zhang, Yu Dept of Electrical and Computer Engineering UC Santa Cruz Dept of Computer Science and Engineering UC Santa Cruz Santa Cruz High School

Encoder-decoder-based recurrent neural network (RNN) has made significant progress in sequence-to-sequence learning tasks such as machine translation and conversational models. Recent works have shown the advantage of this type of network in dealing with various time series forecasting tasks. The present paper focuses on the problem of multi-horizon short-term load forecasting, which plays a key role in the power system's planning and operation. Leveraging the encoder-decoder RNN, we develop an attention model to select the relevant features and similar temporal information adaptively. First, input features are assigned with different weights by a feature selection attention layer, while the updated historical features are encoded by a bi-directional long short-term memory (BiLSTM) layer. Then, a decoder with hierarchical temporal attention enables a similar day selection, which re-evaluates the importance of historical information at each time step. Numerical results tested on the dataset of the global energy forecasting competition 2014 show that our proposed model significantly outperforms some existing forecasting schemes. Copyright © 2021, The Authors. All rights reserved.

关键词： Recurrent neural networks

Unsupervised enrichment of persona-grounded dialog with background stories

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Majumder, Bodhisattwa Prasad Berg-Kirkpatrick, Taylor McAuley, Julian Jhamtani, Harsh Department of Computer Science and Engineering UC San Diego School of Computer Science Carnegie Mellon University

Humans often refer to personal narratives, life experiences, and events to make a conversation more engaging and rich. While persona-grounded dialog models are able to generate responses that follow a given persona, they often miss out on stating detailed experiences or events related to a persona, often leaving conversations shallow and dull. In this work, we equip dialog models with 'background stories' related to a persona by leveraging fictional narratives from existing story datasets (e.g. ROCStories). Since current dialog datasets do not contain such narratives as responses, we perform an unsupervised adaptation of a retrieved story for generating a dialog response using a gradient-based rewriting technique. Our proposed method encourages the generated response to be fluent (i.e., highly likely) with the dialog history, minimally different from the retrieved story to preserve event ordering and consistent with the original persona. We demonstrate that our method can generate responses that are more diverse, and are rated more engaging and human-like by human evaluators, compared to outputs from existing dialog models. © 2021, CC BY.

关键词：

Incorporating data uncertainty in object tracking algorithms

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Muthali, Anish Laine, Forrest Tomlin, Claire The Department of Electrical Engineering and Computer Sciences UC Berkeley The Faculty of Computer Science Vanderbilt University United States The Faculty of Electrical Engineering and Computer Sciences UC Berkeley

Methodologies for incorporating the uncertainties characteristic of data-driven object detectors into object tracking algorithms are explored. Object tracking methods rely on measurement error models, typically in the form of measurement noise, false positive rates, and missed detection rates. Each of these quantities, in general, can be dependent on object or measurement location. However, for detections generated from neural-network processed camera inputs, these measurement error statistics are not sufficient to represent the primary source of errors, namely a dissimilarity between run-time sensor input and the training data upon which the detector was trained. To this end, we investigate incorporating data uncertainty into object tracking methods such as to improve the ability to track objects, and particularly those which out-of-distribution w.r.t. training data. The proposed methodologies are validated on an object tracking benchmark as well on experiments with a real autonomous aircraft. © 2021, CC BY.

关键词： Object detection

Special Session: CAD for Hardware Security - Promising Directions for Automation of Security Assurance

学校读者我要写书评

暂无评论

Special Session: CAD for Hardware Security - Promising Direc...

VLSI Test Symposium

作者： Sohrab Aftabjahani Mark Tehranipoor Farimah Farahmandi Bulbul Ahmed Ryan Kastner Francesco Restuccia Andres Meza Kaki Ryan Nicole Fern Jasper van Woudenberg Rajesh Velegalati Cees-Bart Breunesse Cynthia Sturton Calvin Deutschbein Data Center Platform Engineering and Architecture Intel Corporation Hillsboro OR USA Department of Electrical and Computer Engineering University of Florida Gainesville FL USA Department of Computer Science and Engineering UC San Diego San Diego CA USA Riscure Inc. San Francisco CA USA Department of Computer Science University of North Carolina Chapel Hill NC USA Department of Computing and Data Science Willamette University Salem OR USA

Hardware security creates a hardware-based security foundation for secure and reliable operation of systems and applications used in our modern life. The presence of design for security, security assurance, and general security design life cycle practices in product life cycle of many large semiconductor design and manufacturing companies these days indicates that the importance of hardware security has been very well observed in industry. However, the high cost, time, and effort for building security into designs and assuring their security - due to using many manual processes - is still an important obstacle for economy of secure product development. This paper presents several promising directions for automation of design for security and security assurance practices to reduce the overall time and cost of secure product development. First, we present security verification challenges of SoCs, possible vulnerabilities that could be introduced inadvertently by tools mapping a design model in one level of abstraction to its lower level, and our solution to the problem by automatically mapping security properties from one level to its lower level incorporating techniques for extension and expansion of the properties. Then, we discuss the foundation necessary for further automation of formal security analysis of a design by incorporating threat model and common security vulnerabilities into an intermediate representation of a hardware model to be used to automatically determine if there is a chance for direct or indirect flow of information to compromise confidentiality or integrity of security assets. Finally, we discuss a pre-silicon-based framework for practical and time-and-cost effective power-side channel leakage analysis, root-causing the side-channel leakage by using the automatically generated leakage profile of circuit nodes, providing insight to mitigate the side-channel leakage by addressing the high leakage nodes, and assuring the effectiveness of the miti

关键词：

GRA-LPO: Graph Convolution Based Leakage Power Optimization

学校读者我要写书评

暂无评论

GRA-LPO: Graph Convolution Based Leakage Power Optimization

Asia and South Pacific Design Automation Conference

作者： Uday Mallappa Chung-Kuan Cheng Electrical and Computer Engineering Computer Science and Engineering UC San Diego La Jolla California USA

ISBN: (数字)9781450379991

ISBN: (纸本)9781728180571

Static power consumption is a critical challenge for IC designs, particularly for mobile and IoT applications. A final post-layout step in modern design flows involves a leakage recovery step that is embedded in signoff static timing analysis tools. The goal of such recovery is to make use of the positive slack (if any) and recover the leakage power by performing cell swaps with footprint compatible variants. Though such swaps result in unaltered routing, the hard constraint is not to introduce any new timing violations. This process can require up to tens of hours of runtime, just before the tapeout, when schedule and resource constraints are tightest. The physical design teams can benefit greatly from a fast predictor of the leakage recovery step: if the eventual recovery will be too small, the entire step can be skipped, and the resources can be allocated elsewhere. If we represent the circuit netlist as a graph with cells as vertices and nets connecting these cells as edges, the leakage recovery step is an optimization step, on this graph. If we can learn these optimizations over several graphs with various logic-cone structures, we can generalize the learning to unseen graphs. Using graph convolution neural networks, we develop a learning-based model, that predicts per-cell recoverable slack, and translate these slack values to equivalent power savings. For designs up to 1.6M instances, our inference step takes less than 12 seconds on a Tesla P100 GPU, and an additional feature extraction, post-processing steps consuming 420 seconds. The model is accurate with relative error under 6.2%, for the design-specific context.

关键词： Schedules Convolution Scalability Predictive models Tools Timing Optimization

Quantum algorithms for sampling log-concave distributions and estimating normalizing constants 22

学校读者我要写书评

暂无评论

Quantum algorithms for sampling log-concave distributions an...

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Andrew M. Childs Tongyang Li Jin-Peng Liu Chunhao Wang Ruizhe Zhang Joint Center for Quantum Information and Computer Science Department of Computer Science and Institute for Advanced Computer Studies University of Maryland Center on Frontiers of Computing Studies and School of Computer Science Peking University Simons Institute and Department of Mathematics UC Berkeley Department of Computer Science and Engineering Pennsylvania State University Department of Computer Science The University of Texas at Austin

ISBN: (纸本)9781713871088

Given a convex function f : ℝd → ℝ, the problem of sampling from a distribution ∝ e-f(x) is called log-concave sampling. This task has wide applications in machine learning, physics, statistics, etc. In this work, we develop quantum algorithms for sampling log-concave distributions and for estimating their normalizing constants $\int_{\mathbb{R}^d}e^{-f(x)}\mathrm{d} x$. First, we use underdamped Langevin diffusion to develop quantum algorithms that match the query complexity (in terms of the condition number κ and dimension d) of analogous classical algorithms that use gradient (first-order) queries, even though the quantum algorithms use only evaluation (zeroth-order) queries. For estimating normalizing constants, these algorithms also achieve quadratic speedup in the multiplicative error e. Second, we develop quantum Metropolis-adjusted Langevin algorithms with query complexity Õ(κ1/2d) and Õ(κ1/2d3/2/ε) for log-concave sampling and normalizing constant estimation, respectively, achieving polynomial speedups in κ, d, ε over the best known classical algorithms by exploiting quantum analogs of the Monte Carlo method and quantum walks. We also prove a 1/ε1-o(1) quantum lower bound for estimating normalizing constants, implying near-optimality of our quantum algorithms in ε.

关键词：

Fast Exact Leverage Score Sampling from Khatri-Rao Products with Applications to Tensor Decomposition

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Bharadwaj, Vivek Malik, Osman Asif Murray, Riley Grigori, Laura Buluç, Aydın Demmel, James Electrical Engineering and Computer Science Department UC Berkeley United States Computational Research Division Lawrence Berkeley National Lab. United States International Computer Science Institute United States Institute of Mathematics EPFL Lab for Simulation and Modelling Paul Scherrer Institute Switzerland

We present a data structure to randomly sample rows from the Khatri-Rao product of several matrices according to the exact distribution of its leverage scores. Our proposed sampler draws each row in time logarithmic in the height of the Khatri-Rao product and quadratic in its column count, with persistent space overhead at most the size of the input matrices. As a result, it tractably draws samples even when the matrices forming the Khatri-Rao product have tens of millions of rows each. When used to sketch the linear least squares problems arising in CANDECOMP / PARAFAC tensor decomposition, our method achieves lower asymptotic complexity per solve than recent state-of-the-art methods. Experiments on billion-scale sparse tensors validate our claims, with our algorithm achieving higher accuracy than competing methods as the decomposition rank grows. © 2023, CC BY.

关键词： Matrix algebra

Joint Diffusion Processes as an Inductive Bias in Sheaf Neural Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Caralt, Ferran Hernandez Gil, Guillermo Bernárdez Duta, Iulia Liò, Pietro Cot, Eduard Alarcón FME FIB and ETSETB Universitat Politècnica de Catalunya-BarcelonaTech Barcelona Spain Department of Electrical and Computer Engineering UC Santa Barbara Santa Barbara United States Department of Computer Science and Technology University of Cambridge Cambridge United Kingdom Department of Electronic Engineering Universitat Politècnica de Catalunya-BarcelonaTech Barcelona Spain

Sheaf Neural Networks (SNNs) naturally extend Graph Neural Networks (GNNs) by endowing a cellular sheaf over the graph, equipping nodes and edges with vector spaces and defining linear mappings between them. While the attached geometric structure has proven to be useful in analyzing heterophily and oversmoothing, so far the methods by which the sheaf is computed do not always guarantee a good performance in such settings. In this work, drawing inspiration from opinion dynamics concepts, we propose two novel sheaf learning approaches that (i) provide a more intuitive understanding of the involved structure maps, (ii) introduce a useful inductive bias for heterophily and oversmoothing, and (iii) infer the sheaf in a way that does not scale with the number of features, thus using fewer learnable parameters than existing methods. In our evaluation, we show the limitations of the real-world benchmarks used so far on SNNs, and design a new synthetic task –leveraging the symmetries of n-dimensional ellipsoids– that enables us to better assess the strengths and weaknesses of sheaf-based models. Our extensive experimentation on these novel datasets reveals valuable insights into the scenarios and contexts where SNNs in general –and our proposed approaches in particular– can be beneficial. © 2024, CC BY-NC-ND.

关键词： Vector spaces

Challenging the Machine: Contestability in Government AI Systems

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Landau, Susan Dempsey, James X. Kamar, Ece Bellovin, Steven M. Pool, Robert Columbia University United States The Cybersecurity and Privacy Center The university's Data Science Institute United States Columbia Law School United States UC Berkeley Law School United States The Department of Computer Science and Engineering The University of Washington United States The Fletcher School The School of Engineering Tufts University United States

In an October 2023 executive order (EO), President Biden issued a detailedbut largely aspirational road map for the safe and responsible development anduse of artificial intelligence (AI). The challenge for the January 24-25, 2024workshop was to transform those aspirations regarding one specific but crucialissue - the ability of individuals to challenge government decisions madeabout themselves - into actionable guidance enabling agencies to develop,procure, and use genuinely contestable advanced automated decision-makingsystems. While the Administration has taken important steps since the October2023 EO, the insights garnered from our workshop remain highly relevant, as therequirements for contestability of advanced decision-making systems are not yetfully defined or implemented. The workshop brought together technologists, members of government agenciesand civil society organizations, litigators, and researchers in an intensivetwo-day meeting that examined the challenges that users, developers, andagencies faced in enabling contestability in light of advanced automateddecision-making systems. To ensure a free and open flow of discussion, themeeting was held under a modified version of the Chatham House *** were free to use any information or details that they learned, butthey may not attribute any remarks made at the meeting by the identity or theaffiliation of the speaker. Thus, the workshop summary that follows anonymizesspeakers and their affiliation. Where an identification of an agency, company,or organization is made, it is done from a public, identified resource and doesnot necessarily reflect statements made by participants at the workshop. This document is a report of that workshop, along with recommendations andexplanatory material. © 2024, CC BY-NC-SA.

关键词： Decision making