检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Liu, Sheng Ye, Haotian Xing, Lei Zou, James Department of Biomedical Data Science Stanford University United States Department of Computer Science Stanford University United States Department of Radiation Oncology Stanford University United States

Large language models (LLMs) demonstrate emergent in-context learning capabilities, where they adapt to new tasks based on example demonstrations. However, in-context learning has seen limited effectiveness in many settings, is difficult to quantitatively control and takes up context window space. To overcome these limitations, we propose an alternative approach that recasts in-context learning as in-context vectors (ICV). Using ICV has two steps. We first use a forward pass on demonstration examples to create the in-context vector from the latent embedding of the LLM. This vector captures essential information about the intended task. On a new query, instead of adding demonstrations to the prompt, we shift the latent states of the LLM using the ICV. The ICV approach has several benefits: 1) it enables the LLM to more effectively follow the demonstration examples;2) it's easy to control by adjusting the magnitude of the ICV;3) it reduces the length of the prompt by removing the in-context demonstrations;4) ICV is computationally much more efficient than fine-tuning. We demonstrate that ICV achieves better performance compared to standard in-context learning and fine-tuning on diverse tasks including safety, style transfer, role-playing and formatting. Moreover, we show that we can flexibly teach LLM to simultaneously follow different types of instructions by simple vector arithmetics on the corresponding ICVs. Code is available at https://***/shengliu66/ICV. Copyright 2024 by the author(s)

关键词： Vectors

来源：评论

学校读者我要写书评

暂无评论

Oja's Algorithm for Streaming Sparse PCA 38

Oja's Algorithm for Streaming Sparse PCA

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Kumar, Syamantak Sarkar, Purnamrita Department of Computer Science UT Austin United States Department of Statistics and Data Sciences UT Austin United States

Oja's algorithm for Streaming Principal Component Analysis (PCA) for n datapoints in a d dimensional space achieves the same sin-squared error O(reff/n) as the offline algorithm in O(d) space and O(nd) time and a single pass through the datapoints. Here reff is the effective rank (ratio of the trace and the principal eigenvalue of the population covariance matrix Σ). Under this computational budget, we consider the problem of sparse PCA, where the principal eigenvector of Σ is s-sparse, and reff can be large. In this setting, to our knowledge, there are no known single-pass algorithms that achieve the minimax error bound in O(d) space and O(nd) time without either requiring strong initialization conditions or assuming further structure (e.g., spiked) of the covariance matrix. We show that a simple single-pass procedure that thresholds the output of Oja's algorithm (the Oja vector) can achieve the minimax error bound under some regularity conditions in O(d) space and O(nd) time. We present a nontrivial and novel analysis of the entries of the unnormalized Oja vector, which involves the projection of a product of independent random matrices on a random initial vector. This is completely different from previous analyses of Oja's algorithm and matrix products, which have been done when the reff is bounded. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The Complexity of Learning LTL, CTL and ATL Formulas 42

The Complexity of Learning LTL, CTL and ATL Formulas

引用

42nd International Symposium on Theoretical Aspects of computer science, STACS 2025

作者： Bordais, Benjamin Neider, Daniel Roy, Rajarshi TU Dortmund University Center for Trustworthy Data Science and Security University Alliance Ruhr Dortmund Germany Department of Computer Science University of Oxford United Kingdom

ISBN: (纸本)9783959773652

We consider the problem of learning temporal logic formulas from examples of system behavior. Learning temporal properties has crystallized as an effective means to explain complex temporal behaviors. Several efficient algorithms have been designed for learning temporal formulas. However, the theoretical understanding of the complexity of the learning decision problems remains largely unexplored. To address this, we study the complexity of the passive learning problems of three prominent temporal logics, Linear Temporal Logic (LTL), Computation Tree Logic (CTL) and Alternating-time Temporal Logic (ATL) and several of their fragments. We show that learning formulas with unbounded occurrences of binary operators is NP-complete for all of these logics. On the other hand, when investigating the complexity of learning formulas with bounded occurrences of binary operators, we exhibit discrepancies between the complexity of learning LTL, CTL and ATL formulas (with a varying number of agents). © Benjamin Bordais, Daniel Neider, and Rajarshi Roy.

关键词： Temporal logic

来源：评论

学校读者我要写书评

暂无评论

Unifying Genetics and Imaging: MRI-Based Classification of MGMT Genetic Subtypes Using Visual Transformers 18

Unifying Genetics and Imaging: MRI-Based Classification of M...

引用

18th IEEE International Conference on Emerging Technologies, ICET 2023

作者： Babar, Ibrar Tahir, Asjid Haider, Arslan Ur Rehman, Hafeez Alazab, Moutaz National University of Computer and Emerging Sciences Department of Computer Science Peshawar Pakistan Oryx Universal College Department of Computing and Data Science Doha Qatar

ISBN: (纸本)9798350328172

Brain cancer can take many forms, but glioblastoma (GBM) is one of the most aggressive. To treat it effectively, doctors need to know the genetic subtype of a specific part of the tumor called the O-6-methylguanine-DNA-methyltransferase (MGMT) promoter. This information is critical for making treatment decisions and predicting outcomes. However, traditional methods for predicting MGMT status have significant limitations. One of the main reasons for this poor performance is that these methods fail to capture small variations in MRI signatures that are essential for accurately predicting MGMT promoter methylation status. In this work, we propose a two-part approach to predict the MGMT methylation status. First, we use autoencoders to extract features from the MRI images, capturing intricate spatial patterns and representations. Then, we feed these encoded features into a visual transformer architecture to learn long-range dependencies and contextually significant relationships among image regions. This innovative fusion of autoencoders and visual transformers allows us to harness both local and global information present in the MRI data. We utilized a dataset consisting of 1010 3D MRI images from patients with GBM. Half of the patients had MGMT promoter methylation, while the other half did not. The model achieved impressive results, including an overall accuracy score of 0.904, recall score of 0.902, precision score of 0.914, F1 score of 0.906, and specificity score of 0.906. These metrics far surpass the accuracy of conventional imaging techniques. © 2023 IEEE.

关键词： Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

Conversation Sentiment Analysis in League of Legends Game Community

Conversation Sentiment Analysis in League of Legends Game Co...

引用

2024 IEEE International Conference on Cybernetics and Innovations, ICCI 2024

作者： Pongkhan, Thutchaphong Khoosirirat, Phutanate Songmuang, Pokpong Kongkachandra, Rachada Thammasat University Data Science And Innovation College Of Interdisciplinary Studies Pathum Thani Thailand Thammasat University Faculty Of Science And Technology Department Of Computer Science Pathum Thani Thailand

ISBN: (纸本)9798350370096

Gaming communities are groups of people from different nations, religions, genders, and ages who come to play games and have discussions about the games. However, some gaming communities are plagued by toxic behavior of people in communities, which negatively affects to other players. A Conversation Sentiment Analysis in the League of Legends Game Community is about analyzing the toxicity of the League of Legends game community by using the Sentiment Analysis methods which have Bag-of-Words and Weight Score method, and then comparing it with three machine learning techniques: Logistic Regression, Support Vector Machine (SVM), and CatBoost, and showing how members in this community have interacted with other members in the community by using the Conversation Mapping Interactive method. © 2024 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Zero-shot causal learning 37

Zero-shot causal learning

引用

37th Conference on Neural Information Processing Systems, NeurIPS 2023

作者： Nilforoshan, Hamed Moor, Michael Roohani, Yusuf Chen, Yining Šurina, Anja Yasunaga, Michihiro Oblak, Sara Leskovec, Jure Department of Computer Science Stanford University United States Department of Biomedical Data Science Stanford University United States School of Computer and Communication Sciences EPFL Switzerland Department of Computer Science University of Ljubljana Slovenia

ISBN: (纸本)9781713899921

Predicting how different interventions will causally affect a specific individual is important in a variety of domains such as personalized medicine, public policy, and online marketing. There are a large number of methods to predict the effect of an existing intervention based on historical data from individuals who received it. However, in many settings it is important to predict the effects of novel interventions (e.g., a newly invented drug), which these methods do not address. Here, we consider zero-shot causal learning: predicting the personalized effects of a novel intervention. We propose CaML, a causal meta-learning framework which formulates the personalized prediction of each intervention's effect as a task. CaML trains a single meta-model across thousands of tasks, each constructed by sampling an intervention, its recipients, and its nonrecipients. By leveraging both intervention information (e.g., a drug's attributes) and individual features (e.g., a patient's history), CaML is able to predict the personalized effects of novel interventions that do not exist at the time of training. Experimental results on real world datasets in large-scale medical claims and cell-line perturbations demonstrate the effectiveness of our approach. Most strikingly, CaML's zero-shot predictions outperform even strong baselines trained directly on data from the test interventions. © 2023 Neural information processing systems foundation. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

LegoDroid: flexible Android app decomposition and instant installation

引用

science China(Information sciences) 2023年第4期66卷 155-173页

作者： Yi LIU Yun MA Xusheng XIAO Tao XIE Xuanzhe LIU School of Computer Science Peking University Key Laboratory of High Confidence Software Technologies(Peking University) Ministry of Education Institute for Artificial Intelligence Peking University Department of Computer and Data Sciences Case Western Reserve University

Current mobile applications(apps) have become increasingly complicated with increasing features that are represented on the graphical user interface associated with application programming interfaces(APIs) to access backend functionality and data. Meanwhile, apps suffer from the “software bloat” in volume. Some app features may be redundant, with respect to those features(from other apps) that the users already desirably and frequently use. However, the current app release model forces users to download and install a full-size installation package rather than optionally choosing only their desired features. Large-size apps can not only increase the local resource consumption, such as CPU, memory, and energy, but also inevitably compromise the user experience, such as the slow load time in the app. In this article, we first conduct an empirical study to characterize the app feature usage when users interact with Android apps,and surprisingly find that users access only a very small subset of app features. Based on these findings,we design a new approach named Lego Droid, which automatically decomposes an Android app for flexible loading and installation, while preserving the expected functionality with a fast and instant app load. With such a method, a slimmer bundle will be downloaded and host the target APIs inside the original app to satisfy users' requirements. We implement a system for Lego Droid and evaluate it with 1000 real-world Android apps. Compared to the original full-size apps, apps optimized by Lego Droid can substantially improve the load time by reducing the base bundle and feature bundles by 13.06% and 10.93%, respectively,along with the app-package installation size by 44.17%. In addition, we also demonstrate that Lego Droid is quite practical with evolving versions, as it can produce substantial reusable code to alleviate the developers' efforts when releasing new app versions.

关键词： performance software bloat instant installation mobile applications program analysis

来源：评论

学校读者我要写书评

暂无评论

SkinMate: Mobile-Based Application for Detecting Multi-Class Skin Diseases Classification Using Pre-Trained MobileNetV2 on CNN Architecture 9

SkinMate: Mobile-Based Application for Detecting Multi-Class...

引用

9th IEEE Information Technology International Seminar, ITIS 2023

作者： Fahrudin, Tresna Maulana Illah, Ibnu Zahy'Atha Universitas Pembangunan Nasional 'Veteran' Jawa Timur Faculty of Computer Science Department of Data Science Surabaya Indonesia

ISBN: (纸本)9798350306835

The skin has a very significant role in humans. But unfortunately, the skin is one of the organs vulnerable to disease. Skin diseases can have an impact on the annoying itching, pain, emotional, and social feelings of the sufferer. The causes of skin diseases include bacteria, fungi, viruses, dust, and allergies. Delay in treatment can exacerbate the condition of the affected skin. Therefore, the research proposed a skin disease detection mobile application for early diagnosis. The total of skin disease images is 2.747 data, which is divided into five classes and separated into training and validation data. The category of the dataset includes Acne Vulgaris, Actinic Keratosis, Nail Fungus, Psoriasis, and Seborrheic Keratoses. The development of this application uses several augmentation techniques so that the model gets input training data from various image forms, and then uses pre-trained MobileNetV2, which has low latency and low power as needed on mobile devices. The development of this application also uses the TensorFlow Lite model for inference with ML Kit on Android. The model performance gave the best accuracy of 90.62%, precision of 89%, recall of 90%, and F1- score of 89%. These results indicate that the classifying of skin diseases using the pre-trained MobileNetV2 on CNN architecture is feasible. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Using Decentralized Identifiers and InterPlanetary File System to Create a Recoverable Rare Disease Patient Identity Framework 23

Using Decentralized Identifiers and InterPlanetary File Syst...

引用

7th International Conference on Medical and Health Informatics, ICMHI 2023

作者： McHale, Sean Murr, Lincoln Zhang, Peng Department of Computer Science Vanderbilt University United States Data Science Institute Vanderbilt University United States

ISBN: (纸本)9798400700712

This paper presents a novel framework for creating a recoverable rare disease patient identity system using blockchain and smart contracts, decentralized identifiers (DIDs), and the InterPlanetary File System (IPFS). Smart contracts are executable code that can be written into decentralized storage such as blockchains to enable tamper-proof data transactions. DIDs provide a secure, decentralized, and extensible way to create, store, and manage digital identities, while IPFS provides a distributed, immutable, and secure storage system for patient identities. Utilizing these technologies with smart contracts, we created a framework to store persistent medical records of patients. Smart contracts additionally allow account recovery without the use of any centralized authority. The framework provides a standardized identity layer that ensures healthcare providers have correct and secure access to a patient's data. The paper explores the advantages of using a decentralized identity system and highlights the potential of this approach to improve the security and universality of medical records for patients with rare diseases. © 2023 ACM.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Analyzing Jordanian Higher Education Instructors' Perception Towards Metaverse Based Education 2

Analyzing Jordanian Higher Education Instructors' Perception...

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Omar, Khalil Abu-Aishah, Saif Matar, Izzeddin Al Smadi, Duha Fakhouri, Hussam Abuarqoub, Mohammad University Of Petra Computer Science Department Amman Jordan University Of Petra Information Security Department Amman Jordan University Of Petra Data Science And Ai Department Amman Jordan

ISBN: (纸本)9798350394962

In recent years, many researchers have become interested in how the Metaverse can be applied in higher education. Although the use of Metaverse in education is still in its early stages, ongoing research on virtual worlds in education provides an opportunity to consider potential challenges in exploring the integration of Metaverse in education. Furthermore, there is a lack of empirical studies exploring the factors that impact the acceptance and use of immersive technologies, such as Augmented Reality, Virtual Reality, and Metaverse in higher education, particularly in developing countries such as Jordan. Therefore, the present study aims to create a comprehensive understanding of the current situation, perceptions, needs, and concerns related to integrating the Metaverse into Jordanian higher education institutes. © 2024 IEEE.

关键词： Augmented reality

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：