检索结果-内蒙古大学图书馆

Enhancing Autonomous Vehicle Perception in Adverse Weather through Image Augmentation during Semantic Segmentation Training

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Kou, Ethan Curran, Noah Henry M. Gunn High School Palo AltoCA United States Electrical Engineering and Computer Sciences UC Berkeley United States Computer Science and Engineering University of Michigan United States

Robust perception is crucial in autonomous vehicle navigation and localization. Visual processing tasks, like semantic segmentation, should work in varying weather conditions and during different times of day. Semantic segmentation is where each pixel is assigned a class, which is useful for locating overall features (1). Training a segmentation model requires large amounts of data, and the labeling process for segmentation data is especially tedious. Additionally, many large datasets include only images taken in clear weather. This is a problem because training a model exclusively on clear weather data hinders performance in adverse weather conditions like fog or rain. We hypothesize that given a dataset of only clear days images, applying image augmentation (such as random rain, fog, and brightness) during training allows for domain adaptation to diverse weather conditions. We used CARLA, a 3D realistic autonomous vehicle simulator, to collect 1200 images in clear weather composed of 29 classes from 10 different towns (2). We also collected 1200 images of random weather effects. We trained encoder-decoder UNet models to perform semantic segmentation. Applying augmentations significantly improved segmentation under weathered night conditions (p © 2024, CC BY-NC-ND.

关键词： Semantic Segmentation

Binding in hippocampal-entorhinal circuits enables compositionality in cognitive maps 38

学校读者我要写书评

暂无评论

Binding in hippocampal-entorhinal circuits enables compositi...

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Kymn, Christopher J. Mazelet, Sonia Thomas, Anthony Kleyko, Denis Frady, E. Paxon Sommer, Friedrich T. Olshausen, Bruno A. Redwood Center for Theoretical Neuroscience Helen Wills Neuroscience Institute UC Berkeley Berkeley United States Université Paris-Saclay ENS Paris-Saclay Gif-sur-Yvette France Department of Electrical and Computer Engineering UC Davis Davis United States Centre for Applied Autonomous Sensor Systems Örebro University Örebro Sweden Intelligent Systems Lab Research Institutes of Sweden Kista Sweden Intel Labs Santa Clara United States Herbert Wertheim School of Optometry & Vision Science UC Berkeley Berkeley United States

We propose a normative model for spatial representation in the hippocampal formation that combines optimality principles, such as maximizing coding range and spatial information per neuron, with an algebraic framework for computing in distributed representation. Spatial position is encoded in a residue number system, with individual residues represented by high-dimensional, complex-valued vectors. These are composed into a single vector representing position by a similarity-preserving, conjunctive vector-binding operation. Self-consistency between the representations of the overall position and of the individual residues is enforced by a modular attractor network whose modules correspond to the grid cell modules in entorhinal cortex. The vector binding operation can also associate different contexts to spatial representations, yielding a model for entorhinal cortex and hippocampus. We show that the model achieves normative desiderata including superlinear scaling of patterns with dimension, robust error correction, and hexagonal, carry-free encoding of spatial position. These properties in turn enable robust path integration and association with sensory inputs. More generally, the model formalizes how compositional computations could occur in the hippocampal formation and leads to testable experimental predictions. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

On the Statistical Complexity for Offline and Low-Adaptive Reinforcement Learning with Structures

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Yin, Ming Wang, Mengdi Wang, Yu-Xiang The Department of Electrical and Computer Engineering Princeton University United States The Halıcıoğlu Data Science Institute UC San Diego United States

This article reviews the recent advances on the statistical foundation of reinforcement learning (RL) in the offline and low-adaptive settings. We will start by arguing why offline RL is the appropriate model for almost any real-life ML problems, even if they have nothing to do with the recent AI breakthroughs that use RL. Then we will zoom into two fundamental problems of offline RL: offline policy evaluation (OPE) and offline policy learning (OPL). It may be surprising to people that tight bounds for these problems were not known even for tabular and linear cases until recently. We delineate the differences between worst-case minimax bounds and instance-dependent bounds. We also cover key algorithmic ideas and proof techniques behind near-optimal instance-dependent methods in OPE and OPL. Finally, we discuss the limitations of offline RL and review a burgeoning problem of low-adaptive exploration which addresses these limitations by providing a sweet middle ground between offline and online RL. Copyright © 2025, The Authors. All rights reserved.

关键词： Adversarial machine learning

A Novel Physical Unclonable Function Based on Hybrid Current Mirror

学校读者我要写书评

暂无评论

Journal of Hardware and Systems Security 2023年第4期7卷 125-137页

作者： C. G., Gisha Chakraborty, Ananda Sankar Chakraborty, Rajat Subhra Antony Jose, Bijoy Mathew, Jimson Department of Electronics Cochin University of Science and Technology Kochi India Berkeley Wireless Research Center UC Berkeley Berkeley USA Department of Computer Science and Engineering IIT Kharagpur Kharagpur India Department of Computer Science Cochin University of Science and Technology Kochi India Department of Computer Science IIT Patna Patna India

Intrinsic manufacturing process variations are extensive, unpredictable, and inevitable in modern semiconductor technology. This is advantageously utilized in physically unclonable function (PUF) circuits for enhancing hardware and software security. Among widely-known PUF variants, the Arbiter PUF based on digital building blocks has a regular structure and low hardware footprint; however, it is susceptible to machine learning-based modeling attacks. In this work, we propose a new low-power and reliable PUF, based on the hybrid current mirror inverter (CMI). The proposed PUF circuit exploits the process variation-induced randomness in the CMI circuits to generate instance-specific challenge-response characteristics. The current mirror ensures the stability of the overdrive voltages across all the transistors in the CMI blocks, thereby improving the reliability of the proposed PUF. The PUF circuit was simulated at a 45 nm CMOS technology node, and its major performance metrics were evaluated. The simulation results demonstrated that the proposed PUF has excellent performance metrics while having low hardware and resource footprint. In addition, the proposed PUF demonstrated robustness against machine learning-based model-building attacks.

关键词：

Task Generalization With AutoRegressive Compositional Structure: Can Learning From D Tasks Generalize to DT Tasks?

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Abedsoltan, Amirhesam Zhang, Huaqing Wen, Kaiyue Lin, Hongzhou Zhang, Jingzhao Belkin, Mikhail Department of Computer Science and Engineering UC San Diego United States Halicioglu Data Science Institute UC San Diego United States Institute for Interdisciplinary Information Sciences Tsinghua University China Stanford University United States Amazon. United States

Large language models (LLMs) exhibit remarkable task generalization, solving tasks they were never explicitly trained on with only a few demonstrations. This raises a fundamental question: When can learning from a small set of tasks generalize to a large task family? In this paper, we investigate task generalization through the lens of AutoRegressive Compositional (ARC) structure, where each task is a composition of T operations, and each operation is among a finite family of D subtasks. This yields a total class of size DT. We first show that generalization to all DT tasks is theoretically achievable by training on only ÕpDq tasks. Empirically, we demonstrate that Transformers achieve such exponential task generalization on sparse parity functions via in-context learning (ICL) and Chain-of-Thought (CoT) reasoning. We further demonstrate this generalization in arithmetic and language translation, extending beyond parity functions. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning

Statistical Barriers to Affine-equivariant Estimation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Zihao Cherapanamjeri, Yeshwanth Department of Electrical Engineering and Computer Science Uc Berkeley United States Csail Massachusetts Institute of Technology United States

We investigate the quantitative performance of affine-equivariant estimators for robust mean estimation. As a natural stability requirement, the construction of such affine-equivariant estimators has been extensively studied in the statistics literature. We quantitatively evaluate these estimators under two outlier models which have been the subject of much recent work: the heavy-tailed and adversarial corruption settings. We establish lower bounds which show that affine-equivariance induces a strict degradation in recovery error with quantitative rates degrading by a factor of √d in both settings. We find that classical estimators such as the Tukey median (Tukey '75) and Stahel-Donoho estimator (Stahel '81 and Donoho '82) are either quantitatively sub-optimal even within the class of affine-equivariant estimators or lack any quantitative guarantees. On the other hand, recent estimators with strong quantitative guarantees are not affine-equivariant or require additional distributional assumptions to achieve it. We remedy this by constructing a new affine-equivariant estimator which nearly matches our lower bound. Our estimator is based on a novel notion of a high-dimensional median which may be of independent interest. Notably, our results are applicable more broadly to any estimator whose performance is evaluated in the Mahalanobis norm which, for affine-equivariant estimators, corresponds to an evaluation in Euclidean norm on isotropic distributions. © 2023, CC BY.

关键词：

Design and Development of a LMS with MVVM Architecture

学校读者我要写书评

暂无评论

Design and Development of a LMS with MVVM Architecture

Sustainable Emerging Innovations in engineering and Technology (ICSEIET), International Conference on

作者： Anurag Singh Negi Himadri Vaidya Rahul Chauhan School of Computing Science & Engineering Galgotias University Greater Noida India School of Computer Science & Engineering Galgotias University Greater Noida India School of Computer Science Institute of Management Studies UC Campus Ghaziabad India

A learning management system promotes elearning and can also be used in open distance e-learning. In recent times, the number of mobile devices being used by peers has increased significantly, and an LMS app has thus gained increased importance in the field of E-learning and distance learning. Moodle is another application that is being used in universities, and in this paper, there is an app based on learning management as Moodle is more on the web side, and this app will be based on MVVM architecture, which is Model View. This paper will thus provide insights into the design and development of LMS app and provide an extensive survey of development process.

关键词：

Beyond black-box advice: learning-augmented algorithms for MDPs with Q-value predictions 23

学校读者我要写书评

暂无评论

Beyond black-box advice: learning-augmented algorithms for M...

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Tongxin Li Yiheng Lin Shaolei Ren Adam Wierman School of Data Science CUHK-SZ China Computing + Mathematical Sciences Caltech Electrical & Computer Engineering UC Riverside

We study the tradeoff between consistency and robustness in the context of a single-trajectory time-varying Markov Decision Process (MDP) with untrusted machine-learned advice. Our work departs from the typical approach of treating advice as coming from black-box sources by instead considering a setting where additional information about how the advice is generated is available. We prove a first-of-its-kind consistency and robustness tradeoff given Q-value advice under a general MDP model that includes both continuous and discrete state/action spaces. Our results highlight that utilizing Q-value advice enables dynamic pursuit of the better of machine-learned advice and a robust baseline, thus result in near-optimal performance guarantees, which provably improves what can be obtained solely with black-box advice.

关键词：

Aggregate and conquer: detecting and steering LLM concepts by combining nonlinear predictors over multiple layers

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Beaglehole, Daniel Radhakrishnan, Adityanarayanan Boix-Adserà, Enric Belkin, Mikhail Computer Science and Engineering UC San Diego United States Broad Institute of MIT Harvard Harvard SEAS United States MIT Mathematics Harvard CMSA United States Halıcıoğlu Data Science Institute UC San Diego United States

A trained Large Language Model (LLM) contains much of human knowledge. Yet, it is difficult to gauge the extent or accuracy of that knowledge, as LLMs do not always "know what they know" and may even be actively misleading. In this work, we give a general method for detecting semantic concepts in the internal activations of LLMs. Furthermore, we show that our methodology can be easily adapted to steer LLMs toward desirable outputs. Our innovations are the following: (1) we use a nonlinear feature learning method to identify important linear directions for predicting concepts from each layer;(2) we aggregate features across layers to build powerful concept detectors and steering mechanisms. We showcase the power of our approach by attaining state-of-the-art results for detecting hallucinations, harmfulness, toxicity, and untruthful content on seven benchmarks. We highlight the generality of our approach by steering LLMs towards new concepts that, to the best of our knowledge, have not been previously considered in the literature, including: semantic disambiguation, human languages, programming languages, hallucinated responses, science subjects, poetic/Shakespearean English, and even multiple concepts simultaneously. Moreover, our method can steer concepts with numerical attributes such as product reviews. We provide our code (including a simple API for our methods) at https://***/dmbeaglehole/neural_***. © 2025, CC0.

关键词： Aggregates