检索结果-内蒙古大学图书馆

Reinforcement learning of implicit and explicit control flow in instructions

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Brooks, Ethan A. Rajendran, Janarthanan Lewis, Richard L. Singh, Satinder Department of Computer Science University of Michigan Weinberg Institute for Cognitive Science Department of Psychology and Linguistics University of Michigan

Learning to flexibly follow task instructions in dynamic environments poses interesting challenges for reinforcement learning agents. We focus here on the problem of learning control flow that deviates from a strict step-by-step execution of instructions-that is, control flow that may skip forward over parts of the instructions or return backward to previously completed or skipped steps. Demand for such flexible control arises in two fundamental ways: explicitly when control is specified in the instructions themselves (such as conditional branching and looping) and implicitly when stochastic environment dynamics require re-completion of instructions whose effects have been perturbed, or opportunistic skipping of instructions whose effects are already present. We formulate an attention-based architecture that meets these challenges by learning, from task reward only, to flexibly attend to and condition behavior on an internal encoding of the instructions. We test the architecture's ability to learn both explicit and implicit control in two illustrative domains-one inspired by Minecraft and the other by StarCraft-and show that the architecture exhibits zero-shot generalization to novel instructions of length greater than those in a training set, at a performance level unmatched by three baseline recurrent architectures and one ablation architecture. © 2021, CC BY.

关键词： Reinforcement learning

ScatTR: Estimating the Size of Long Tandem Repeat Expansions from Short-Reads 29th

学校读者我要写书评

暂无评论

ScatTR: Estimating the Size of Long Tandem Repeat Expansio...

29th International Conference on Research in Computational Molecular Biology, RECOMB 2025

作者： Al-Abri, Rashid Gürsoy, Gamze Department of Biomedical Informatics Columbia University New YorkNY10032 United States New York Genome Center New YorkNY10013 United States Department of Computer Science Columbia University New YorkNY10027 United States

ISBN: (纸本)9783031902512

Tandem repeats (TRs) are sequences of DNA where two or more base pairs are repeated back-to-back at specific locations in the genome. The expansions of TRs are implicated in over 50 conditions, including Friedreich’s ataxia, autism, and cancer. However, accurately measuring the copy number of TRs is challenging, especially when their expansions are larger than the fragment sizes used in standard short-read genome sequencing. Here we introduce ScatTR, a novel computational method that leverages a maximum likelihood framework to estimate the copy number of large TR expansions from short-read sequencing data. ScatTR calculates the likelihood of different alignments between sequencing reads and reference sequences that represent various TR lengths and employs a Monte Carlo technique to find the best match. In simulated data, ScatTR outperforms state-of-the-art methods, particularly for TRs with longer motifs and those with lengths that greatly exceed typical sequencing fragment sizes. When applied to data from the 1000 Genomes Project, ScatTR detected potential large TR expansions that other methods missed, highlighting its improved ability to identify genome-wide TR variation. ScatTR can be accessed via: https://***/rashidalabri/scattr. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Simulated annealing

Domain Generalization for Medical Image Analysis: A Review

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yoon, Jee Seok Oh, Kwanseok Shin, Yooseung Mazurowski, Maciej A. Suk, Heung-Il Department of Brain and Cognitive Engineering Korea University Seoul02841 Korea Republic of Department of Artificial Intelligence Korea University Seoul02841 Korea Republic of Departments of Radiology Biostatistics and Bioinformatics Electrical and Computer Engineering and Computer Science Duke University DurhamNC27705 United States Department of Artificial Intelligence The Department of Brain and Cognitive Engineering Korea University Seoul02841 Korea Republic of

Medical image analysis (MedIA) has become an essential tool in medicine and healthcare, aiding in disease diagnosis, prognosis, and treatment planning, and recent successes in deep learning (DL) have made significant contributions to its advances. However, deploying DL models for MedIA in real-world situations remains challenging due to their failure to generalize across the distributional gap between training and testing samples — a problem known as domain shift. Researchers have dedicated their efforts to developing various DL methods to adapt and perform robustly on unknown and out-of-distribution data distributions. This paper comprehensively reviews domain generalization studies specifically tailored for MedIA. We provide a holistic view of how domain generalization techniques interact within the broader MedIA system, going beyond methodologies to consider the operational implications on the entire MedIA workflow. Specifically, we categorize domain generalization methods into data-level, feature-level, model-level, and analysis-level methods. We show how those methods can be used in various stages of the MedIA workflow with DL equipped from data acquisition to model prediction and analysis. Furthermore, we critically analyze the strengths and weaknesses of various methods, unveiling future research opportunities. Copyright © 2023, The Authors. All rights reserved.

关键词： Reviews

Rethinking Masked Language Modeling for Chinese Spelling Correction

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Wu, Hongqiu Zhang, Shaohua Zhang, Yuchen Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China ByteDance China

In this paper, we study Chinese Spelling Correction (CSC) as a joint decision made by two separate models: a language model and an error model. Through empirical analysis, we find that fine-tuning BERT tends to over-fit the error model while under-fit the language model, resulting in poor generalization to out-of-distribution error patterns. Given that BERT is the backbone of most CSC models, this phenomenon has a significant negative impact. To address this issue, we are releasing a multi-domain benchmark LEMON, with higher quality and diversity than existing benchmarks, to allow a comprehensive assessment of the open domain generalization of CSC models. Then, we demonstrate that a very simple strategy – randomly masking 20% non-error tokens from the input sequence during fine-tuning – is sufficient for learning a much better language model without sacrificing the error model. This technique can be applied to any model architecture and achieves new state-of-the-art results on SIGHAN, ECSpell, and LEMON Copyright © 2023, The Authors. All rights reserved.

关键词： Errors

Using annotation for computerized support for fast skimming of cardiology electronic health record notes

学校读者我要写书评

暂无评论

Using annotation for computerized support for fast skimming ...

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Mahshad Koohi H. Dehkordi Andrew J. Einstein Shuxin Zhou Gai Elhanan Yehoshua Perl Vipina K. Keloth James Geller Hao Liu Department of Computer Science NJIT Newark NJ USA Dept. of Medicine Cardiology Division Irving Medical Center New York NY USA Renown Institute for Health Innovation Desert Research Institute Reno NV USA Biomedical Informatics and Data Science Yale School of Medicine New Haven CT USA Department of Data Science NJIT Newark NJ USA Department of Computer Science Montclair State University Montclair NJ USA

Under the circumstances prevalent in current healthcare, medical professionals such as physicians and nurses need to read large numbers of Electronic Health Record (EHR) notes. This demand fosters a situation in which providers typically do not read a whole note but quickly skim it, to capture its essential content. Consequently, by fast skimming, one may miss a critical medical fact. Annotation highlights important content in EHR notes, and enables healthcare professionals to perform fast skimming, thereby minimizing the risk of missing critical information, which is detrimental to patient care. We designed the Cardiology Interface Terminology (CIT) for the purpose of annotation of cardiology EHRs. We emphasize that by annotation we refer to highlighting important information in the EHR notes for enabling fast skimming rather than just recognizing names of diseases, drugs, etc. from Reference Terminologies as is usually done by existing Named Entity Recognition (NER) systems. The CIT design starts with the cardiology components of SNOMED CT. It is enhanced by mining phrases from cardiology EHRs, as potential CIT concepts, which are of higher granularity than SNOMED concepts. Machine learning (ML), the state of the art technique for mining concepts from EHRs, requires training data. However, there is no training data for designing CIT. In the first stage, we introduce an innovative semi-automatic method for mining concepts from EHRs, to replace costly manual mining. The only manual portion is the review of the automatically mined phrases, before their insertion as CIT concepts. The effectiveness of annotation of cardiology EHRs with CIT was evaluated utilizing proper metrics, and compared to annotation with SNOMED CT. In a future second stage, ML mining techniques will be used for enhancing CIT with extra concepts from EHRs, utilizing the concepts added in the first stage as training data. This work focuses on a novel semi-automated method to design the Cardiology I

关键词：

Tableaux for Automated Reasoning in Dependently-Typed Higher-Order Logic

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Niederhauser, Johannes Brown, Chad E. Kaliszyk, Cezary Department of Computer Science University of Innsbruck Innsbruck Austria Czech Institute of Informatics Robotics and Cybernetics Czech Technical University in Prague Prague Czech Republic School of Computing and Information Systems University of Melbourne Melbourne Australia

Dependent type theory gives an expressive type system facilitating succinct formalizations of mathematical concepts. In practice, it is mainly used for interactive theorem proving with intensional type theories, with PVS being a notable exception. In this paper, we present native rules for automated reasoning in a dependently-typed version (DHOL) of classical higher-order logic (HOL). DHOL has an extensional type theory with an undecidable type checking problem which contains theorem proving. We implemented the inference rules as well as an automatic type checking mode in Lash, a fork of Satallax, the leading tableaux-based prover for HOL. Our method is sound and complete with respect to provability in DHOL. Completeness is guaranteed by the incorporation of a sound and complete translation from DHOL to HOL recently proposed by Rothgang et al. While this translation can already be used as a preprocessing step to any HOL prover, to achieve better performance, our system directly works in DHOL. Moreover, experimental results show that the DHOL version of Lash can outperform all major HOL provers executed on the translation. © 2024, CC BY.

关键词： computer circuits

Enhancing Martian Mineral Identification Using an Artificial Neural Network With Extracted Spectral Features In CRISM MTRDR Data

学校读者我要写书评

暂无评论

Enhancing Martian Mineral Identification Using an Artificial...

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Priyanka Kumari Sampriti Soor Amba Shetty B. S. Daya Sagar Department of Water Resources and Ocean Engineering NIT Karnataka Surathkal India School of Computer Engineering Kalinga Institute of Industrial Technology Bhubaneswar India Systems Science and Informatics Unit Indian Statistical Institute Bangalore India

ISBN: (数字)9798350360325

ISBN: (纸本)9798350360332

Creating a supervised learning model for mineral identification is challenging due to the lack of ground-truth data. This study utilizes a method from existing literature that generates a training dataset by augmenting available spectra in the MICA spectral library. However, rather than using entire spectra for identification, this study extracts spectral features from each spectrum for model training. It employs the apparent continuum removal method, Segmented Curve Fitting, to identify the most informative or distinguishable parts in the spectral domain. Spectral features are then extracted based on band-centers and band-areas for each selected part. The model is evaluated against a Targeted Reduced Data Record (TRDR) dataset obtained using a hierarchical Bayesian model, demonstrating improved identification performance than the existing supervised models. Finally, using this model, dominant minerals are identified in MTRDR data from the Nilli Fossae region of Mars, and a corresponding mineral map is presented.

关键词： Training Absorption Supervised learning Artificial neural networks Feature extraction Data models Libraries

Ask, Attend, Attack: A Effective Decision-Based Black-Box Targeted Attack for Image-to-Text Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zeng, Qingyuan Wang, Zhenzhong Cheung, Yiu-Ming Jiang, Min Institute of Artificial Intelligence Xiamen University Fujian China Department of Computing The Hong Kong Polytechnic University Hong Kong Department of Computer Science Hong Kong Baptist University Hong Kong School of Informatics Xiamen University Fujian China

While image-to-text models have demonstrated significant advancements in various vision-language tasks, they remain susceptible to adversarial attacks. Existing white-box attacks on image-to-text models require access to the architecture, gradients, and parameters of the target model, resulting in low practicality. Although the recently proposed gray-box attacks have improved practicality, they suffer from semantic loss during the training process, which limits their targeted attack performance. To advance adversarial attacks of image-to-text models, this paper focuses on a challenging scenario: decision-based black-box targeted attacks where the attackers only have access to the final output text and aim to perform targeted attacks. Specifically, we formulate the decision-based black-box targeted attack as a large-scale optimization problem. To efficiently solve the optimization problem, a three-stage process Ask, Attend, Attack, called AAA, is proposed to coordinate with the solver. Ask guides attackers to create target texts that satisfy the specific semantics. Attend identifies the crucial regions of the image for attacking, thus reducing the search space for the subsequent Attack. Attack uses an evolutionary algorithm to attack the crucial regions, where the attacks are semantically related to the target texts of Ask, thus achieving targeted attacks without semantic loss. Experimental results on transformer-based and CNN+RNNbased image-to-text models confirmed the effectiveness of our proposed AAA. © 2024, CC BY-NC-SA.

关键词： Visual languages

AI and IoT in Smart Cities: A Methodology, Transformation, and Challenges 8th

学校读者我要写书评

暂无评论

AI and IoT in Smart Cities: A Methodology, Transformation, a...

8th International Conference on Information System Design and Intelligent Applications, ISDIA 2024

作者： Begishev, Ildar Isavnin, Alexey Nedelkin, Alexey Lydia, E. Laxmi Kumar, K. Vijaya Doctor of Law Institute of Digital Technologies and Law Department of Criminal Law and Procedure Kazan Innovative University named after V. G. Timiryasov Kazan Russia Doctor of Physical-Mathematical Sciences Department of Business-Informatics and Mathematical Methods in Economics Kazan Federal University Naberezhnye Chelny Institute KFU Naberezhnye Chelny Russia Department of Computer Science Plekhanov Russian University of Economics Moscow Russia Vijayawada520007 India GITAM School of Technology Visakhapatnam Campus Visakhapatnam India

ISBN: (纸本)9789819748945

The Internet of Things (IoT) with smart cities and smart homes enables us to progressively sense and alter our surroundings. However, Artificial Intelligence (AI) must leverage such distinguishing abilities and detected data. We are heading in that way with sophisticated IoT technologies. As connected items become more authoritative, we are seeing a smart city movement away from cloud-based IoT systems and toward edge AI and embedded AI. Lower latency, better privacy, and the necessity to process data close to the source have all influenced this decision. According to a current report, smart cities can have 70 billion linked things by 2023. These cities will become smarter as a result of these networked things and introduce hazards and privacy concerns. As a result of the numerous smart city initiatives and plans that have been implemented in current years, we will not only see the anticipated aids but also the hazards that have been presented. The present and future movements in smart cities and IoT are discussed. This chapter provides an overview of several smart city techniques and elements that can be or are now automated with the help of the Internet of Things (IoT), and Artificial Intelligence (AI). Novel components of smart cities of smart parking and smart governance are determined through Artificial Intelligence (AI) in the Internet of Things (IoT). Finally, the IoT's flaws and solutions in the context of smart cities are determined. This work focuses on identifying the transformation and challenges faced in developing a smart city in AI. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Artificial intelligence (AI) Challenges Internet of things (IoT) Smart cities Transformation