48 Hits in 5.1 sec

Binary Classification with XOR Queries: Fundamental Limits and An Efficient Algorithm [article]

Daesung Kim, Hye Won Chung
2021 arXiv   pre-print
We consider a query-based data acquisition problem for binary classification of unknown labels, which has diverse applications in communications, crowdsourcing, recommender systems and active learning.  ...  Further, we propose an efficient inference algorithm that achieves this limit even when the noise parameters are unknown.  ...  A more important aspect of mixing queries with different degrees we argue in this paper is that it is possible to achieve the optimal sample complexity (2) with an efficient algorithm if there are Θ(m)  ... 
arXiv:2001.11775v2 fatcat:aslehciuinhgzdrmmn7ihskwgq

Parity Queries for Binary Classification [article]

Hye Won Chung, Ji Oon Lee, Doyeon Kim, Alfred O. Hero
2019 arXiv   pre-print
In particular, the necessary and sufficient sample complexity required for recovering all k variables with high probability is n = c_0 max{k, (k log k)/d̅} and the sample complexity for recovering a fixed  ...  We obtain fundamental trade-offs between recovery accuracy, query difficulty, and sample complexity.  ...  On the other hand, we propose a set of queries (an optimal query degree distribution) with the same average query degreed for which the computationally-efficient BP algorithm can almost exactly recover  ... 
arXiv:1809.00901v2 fatcat:rwl2wudwqfhphhsi32dovzzos4

ALdataset: a benchmark for pool-based active learning [article]

Xueying Zhan, Antoni Bert Chan
2020 arXiv   pre-print
Active learning (AL) is a subfield of machine learning (ML) in which a learning algorithm could achieve good accuracy with less training samples by interactively querying a user/oracle to label new data  ...  selected samples by multiple query criteria using a weighted sum of objectives or other multi-objective optimization methods.  ...  Introduction AL is an effective protocol of supervised ML, which selects the most critical instances and queries their labels through the interaction with oracles (experts or multiple human annotators)  ... 
arXiv:2010.08161v1 fatcat:hpcxi5rfxbcvpoa6uagahgcaum

ImitAL: Learning Active Learning Strategies from Synthetic Data [article]

Julius Gonsior, Maik Thiele, Wolfgang Lehner
2021 arXiv   pre-print
To show the general and superior applicability of , we perform an extensive evaluation comparing our strategy on 15 different datasets, from a wide range of domains, with 10 different state-of-the-art  ...  Active Learning (AL) is a well-known standard method for efficiently obtaining labeled data by first labeling the samples that contain the most information based on a query strategy.  ...  optimal AL query strategy would go straight to an F1-Score of 1.0, and continuing on the top, resulting in a rectangle with an area of 1.  ... 
arXiv:2108.07670v1 fatcat:zxrupyw3yng7xc22pivcg4nd4y

The Computational Limits of Deep Learning [article]

Neil C. Thompson, Kristjan Greenewald, Keeheon Lee, Gabriel F. Manso
2020 arXiv   pre-print
But this progress has come with a voracious appetite for computing power.  ...  The prediction MSE is averaged over query vectors sampled from an isotropic Gaussian distribution.  ...  For example, an image classification algorithm performs a set of instructions to calculates the probability that an image contains a cat or not.  ... 
arXiv:2007.05558v1 fatcat:w2grqtaksjaydk4o64rpegpfxu

Neural Pseudo-Label Optimism for the Bank Loan Problem [article]

Aldo Pacchiano, Shaun Singh, Edward Chou, Alexander C. Berg, Jakob Foerster
2021 arXiv   pre-print
As a result, it is possible for the lender's algorithm to "get stuck" with a self-fulfilling model.  ...  We present Pseudo-Label Optimism (PLOT), a conceptually and computationally simple method for this setting applicable to DNNs. adds an optimistic label to the subset of decision points the current model  ...  There are four clusters in the XOR dataset. Each of these is produced by sampling a multivariate normal with isotropic covariance with a diagonal value of 0.5.  ... 
arXiv:2112.02185v1 fatcat:quqq35cr3rb73n76czvypcgy24

CAN-D: A Modular Four-Step Pipeline for Comprehensively Decoding Controller Area Network Data [article]

Miki E. Verma, Robert A. Bridges, Jordan J. Sosnowski, Samuel C. Hollifield, Michael D. Iannacone
2021 arXiv   pre-print
We formulate, formally analyze, and provide an efficient solution to an optimization problem, allowing identification of the optimal set of signal boundaries and byte orderings.  ...  algorithm which exhibits a >97% F-score.  ...  This is an open, crowdsourced set of DBCs that was constructed by individuals using a CommaAI Panda device (an OBD-II plugin) along with the CommaAI Cabana interface to hand label the data for their vehicle  ... 
arXiv:2006.05993v2 fatcat:l5z3zkf7dvh3dbb7g4issz2d7y

A Survey on Smartphone-Based Crowdsensing Solutions

Willian Zamora, Carlos T. Calafate, Juan-Carlos Cano, Pietro Manzoni
2016 Mobile Information Systems  
These factors have led to an outstanding growth of crowdsensing proposals from both academia and industry.  ...  In recent years, the widespread adoption of mobile phones, combined with the ever-increasing number of sensors that smartphones are equipped with, greatly simplified the generalized adoption of crowdsensing  ...  XOR operation).  ... 
doi:10.1155/2016/9681842 fatcat:zzmclqr72nd5nbloyoo6stgzlq

2020 Index IEEE Transactions on Information Forensics and Security Vol. 15

2020 IEEE Transactions on Information Forensics and Security  
Dai, W., +, TIFS 2020 725-737 Communication complexity A New Capacity-Achieving Private Information Retrieval Scheme With (Almost) Optimal File Length for Coded Servers.  ...  ., +, TIFS 2020 2253-2267 Learning a Compact Vein Discrimination Model With GANerated Samples.  ... 
doi:10.1109/tifs.2021.3053735 fatcat:eforexmnczeqzdj3sc2j4yoige

Textual and Content-Based Search in Repositories of Web Application Models

Bojana Bislimovska, Alessandro Bozzon, Marco Brambilla, Piero Fraternali
2014 ACM Transactions on the Web  
Keyword-based and content-based search (also known as query-by-example) are contrasted, with respect to the architecture of the system, the processing of models and queries, and the way in which metamodel  ...  A thorough experimental evaluation is conducted to examine what parameter configurations lead to better accuracy and to offer an insight in what queries are addressed best by each system.  ...  : a greedy algorithm, an exhaustive algorithm with pruning, a process heuristic algorithm, and the A-star algorithm.  ... 
doi:10.1145/2579991 fatcat:56jhidk23bc3bj6wzsl5sq72ha

Neural Approaches to Conversational AI [article]

Jianfeng Gao, Michel Galley, Lihong Li
2019 arXiv   pre-print
Here, we are concerned with the use of a user model to generate more data to improve sample complexity in optimizing a dialogue system.  ...  More complex queries need more steps.  ... 
arXiv:1809.08267v3 fatcat:j57xlm4ogferdnrpfs4f2jporq

Introduction [chapter]

2016 Music Data Analysis  
A similar "query-by-singing" ap proach was later proposed in [28] , which was, however, not combined with an acous tic retrieval method.  ...  In the general case of a causal parametric LTI system, input samples with indices l ≤ k and output samples with indices l < k contribute to an output sample y[k] as y [k] = b 0 x[k]+b 1 x[k−1]+ · · ·+b  ...  An XOR gate corresponds to the hardware implementation of such an XOR operator that is typically included in standard cell libraries.  ... 
doi:10.1201/9781315370996-5 fatcat:avooqogcpnbjngqmzuonil3exq

Identifying and Exploiting Structures for Reliable Deep Learning [article]

Amartya Sanyal
2021 arXiv   pre-print
Deep learning research has recently witnessed an impressively fast-paced progress in a wide range of tasks including computer vision, natural language processing, and reinforcement learning.  ...  To do this, we identify structures in deep neural networks that can be exploited to mitigate the above causes of unreliability of deep learning algorithms.  ...  Further, when along with classification error the learning problem also requires controlling an additional metric of reliability like adversarial vulnerability or mis-calibration, the sample complexity  ... 
arXiv:2108.07083v1 fatcat:lducrn5tlfeqvpxevz6gukfvse

Password Similarity Using Probabilistic Data Structures

Davide Berardi, Franco Callegati, Andrea Melis, Marco Prandini
2020 Journal of Cybersecurity and Privacy  
This work describes an approach based on Bloom Filters to detect password similarity, which can be used to discourage password reuse habits.  ...  This speed gap was due to the intrinsic complexity of the algorithm.  ...  After this generation step, the encryption stream can be combined with the payload, using a xor operation.  ... 
doi:10.3390/jcp1010005 fatcat:ka5d7iu74vbcdolu7zkyh6j6pa

Binary convolutional neural network features off-the-shelf for image to video linking in endoscopic multimedia databases

Stefan Petscharnig, Klaus Schöffmann
2018 Multimedia tools and applications  
With a rigorous long-term archival of endoscopic surgeries, vast amounts of video and image data accumulate.  ...  For evaluation, we use 5.5 h of endoscopic video material and 69 query images selected by medical experts and compare the performance of the aforementioned image mathing methods in terms of video hit rate  ...  Algorithm 1 shows the straightforward binarization algorithm for an EMDB. Given is a real-valued feature set F with k observations of n real-valued feature components.  ... 
doi:10.1007/s11042-018-6016-3 fatcat:ka5y3sjs7ffftipreutmnelbsu
« Previous Showing results 1 — 15 out of 48 results