Probability and Statistics for Computer Science

Forsyth, David

doi:10.1007/978-3-319-64410-3

Cited by 34 publications

(22 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…4.1) with N = 10, 000 and p I ≈ 30% (the plotted results are the median over 20 runs). Our obtained results conform to statistical theory [24]. Specifically, the inlier rates in the subsets of N sample points have a mean of p I and standard deviation (SD) of p I (1−p I ) Nsample .…”

Section: Ablation Studiessupporting

confidence: 85%

SDRSAC: Semidefinite-Based Randomized Approach for Robust Point Cloud Registration Without Correspondences

Hoang

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

This paper presents a novel randomized algorithm for robust point cloud registration without correspondences. Most existing registration approaches require a set of putative correspondences obtained by extracting invariant descriptors. However, such descriptors could become unreliable in noisy and contaminated settings. In these settings, methods that directly handle input point sets are preferable. Without correspondences, however, conventional randomized techniques require a very large amount of samples in order to reach satisfactory solutions. In this paper, we propose a novel approach to address this problem. In particular, our work enables the use of randomized methods for point cloud registration without the need of putative correspondences. By considering point cloud alignment as a special instance of graph matching and employing an efficient semi-definite relaxation, we propose a novel sampling mechanism, in which the size of the sampled subsets can be larger-than-minimal. Our tight relaxation scheme enables fast rejection of the outliers in the sampled sets, resulting in high quality hypotheses. We conduct extensive experiments to demonstrate that our approach outperforms other state-of-the-art methods. Importantly, our proposed method serves as a generic framework which can be extended to problems with known correspondences 1 .

show abstract

Section: Ablation Studiessupporting

confidence: 85%

SDRSAC: Semidefinite-Based Randomized Approach for Robust Point Cloud Registration Without Correspondences

Hoang

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

show abstract

“…Note that we also identified (and subsequently discarded) an additional "Junk" cluster with relatively fewer elements (9.7% of the total number of connectivity profiles computed for all neurons in all recordings) and small values of all features. We thus removed these cases, as usual in unsupervised clustering applications 44 , to better discriminate the remaining "interesting" classes listed above.…”

Section: /32mentioning

confidence: 99%

Dynamic core-periphery structure of information sharing networks in entorhinal cortex and hippocampus

Pedreschi

Bernard

Clawson

et al. 2020

Network Neuroscience

View full text Add to dashboard Cite

Neural computation is associated with the emergence, reconfiguration and dissolution of cell assemblies in the context of varying oscillatory states. Here, we describe the complex spatio-temporal dynamics of cell assemblies through temporal network formalism. We use a sliding window approach to extract sequences of networks of information sharing among single units in hippocampus and enthorinal cortex during anesthesia and study how global and node-wise functional connectivity properties evolve along time and as a function of changing global brain state (theta vs slow-wave oscillations). First, we find that information sharing networks display, at any time, a core-periphery structure in which an integrated core of more tightly functionally interconnected units link to more loosely connected network leaves. However the units participating to the core or to the periphery substantially change across time-windows, with units entering and leaving the core in a smooth way. Second, we find that discrete network states can be defined on top of this continuously ongoing liquid core-periphery reorganization. Switching between network states results in a more abrupt modification of the units belonging to the core and is only loosely linked to transitions between global oscillatory states. Third, we characterize different styles of temporal connectivity that cells can exhibit within each state of the sharing network. While inhibitory cells tend to be central, we show that, otherwise, anatomical localization only poorly influences the patterns of temporal connectivity of the different cells. Furthermore, cells can change temporal connectivity style when the network changes state. Altogether, these findings reveal that the sharing of information mediated by the intrinsic dynamics of hippocampal and enthorinal cortex cell assemblies have a rich spatiotemporal structure, which could not have been identified by more conventional time-or state-averaged analyses of functional connectivity. arXiv:2001.06371v1 [q-bio.NC] 17 Jan 2020 between individuals 16, 17 , multiple temporal and structural scales [19][20][21] , and a rich array of intrinsically dynamical structures that could not be unveiled within a static network framework [22][23][24][25] . Taking into account temporality has moreover been shown to have a strong impact in processes taking place on networks, in particular the propagation of diseases or of information 18, 26, 27 . In the neuroscience field, emphasis have been recently put on the need to upgrade "connectomics" into "chronnectomics", to disentangle temporal variability from inter-subject and intercohort differences and thus achieve a superior biomarking performance 28, 29 . However a majority of dynamic network studies have been so far considering large-scale brain-wide networks of interregional connectivity --see, e.g. 30 for a study of link burstiness-and only fewer have addressed the dynamics of information sharing networks at the level of micro-circuits, often in vitro or in silico [31][32][33] and even mor...

show abstract

“…MLC trading strategy [35] brings a high level of reliability because it is based on several algorithms that are tested simultaneously: decision trees [50], linear discriminant analysis by Fisher [18], support vector machines [5,19], k-nearest neighbors (k-NN) [1], and ensemble learning methods [45]. Three different branches of decision trees are used in our MLC trading strategy: fine, medium, and coarse, which differ according to their complexity, i.e., the number of splits [26]. Several support vector machines (SVMs) are tested: linear, quadratic, and cubic SVM, as well as fine Gaussian, medium Gaussian, and coarse Gaussian.…”

Section: Machine Learning Classifiersmentioning

confidence: 99%

Deep Learning for Stock Market Trading: A Superior Trading Strategy?

Fister¹,

Mun²,

Jagrić³

et al. 2019

NNW

View full text Add to dashboard Cite

Deep-learning initiatives have vastly changed the analysis of data. Complex networks became accessible to anyone in any research area. In this paper we are proposing a deep-learning long short-term memory network (LSTM) for automated stock trading. A mechanical trading system is used to evaluate its performance. The proposed solution is compared to traditional trading strategies, i.e., passive and rule-based trading strategies, as well as machine learning classifiers. We have discovered that the deep-learning long short-term memory network has outperformed other trading strategies for the German blue-chip stock, BMW, during the 2010-2018 period. system (MTS) to sequentially (day-by-day) execute trading signals (decisions), and thus trade with real stocks [66] in virtual time. Consequently, the quality and effectiveness of a trading system over a period of a few years can be evaluated in a matter of seconds.The MTS works by giving three common trading signals, i.e., buy, hold, and sell, for each of the stocks. It is given an initial amount of cash to buy stocks that can be held in the portfolio or be sold at a later date. Buying and selling stocks can generate profits if trading signals are given rationally or generate loss if they are not. Profits and losses can be monitored and reviewed at the end of the trading period to realize strengths and weaknesses of the trading decisions [47]. Trading decisions are typically given by trading strategies, i.e., automated algorithms that constantly monitor market behaviour and react accordingly. Multiple trading strategies are usually incorporated within the trading system, by each giving unique trading decisions. The latter can be evaluated quickly and inexpensively either in realworld time or using the MTS.In this paper we are proposing a concept of an automated, single stock, trading system using the MTS, examining its potential by five different trading strategies, realizing their strengths and weaknesses, and proving the correctness of an optimal strategy. Based on this, we encourage further analysis and the design of an automated portfolio trading system for many similarly treated stocks in parallel. Here we employ three different trading strategies: (1) passive, (2) rule-based -relative strength index (RSI) and moving average convergence/divergence (MACD) technical indicators, and (3) surrogate model trading strategies using machine learning classifiers (MLC) and long short-term memory network (LSTM). Each of them can provide three trading signals: buy, hold, or sell.Since deep learning (DL) [37] is primarily intended for engineering applications, such as image and sound processing, we have not found many examples of DL in the fields of finance, banking, or insurance. In line with this, we would like to apply the DL to the area of finance, particularly mechanical trading systems as an alternative, and thus test whether the DL can be successfully deployed into this area.The structure of the paper is as follows: Section 2 presents the literature review, MLC application...

show abstract

Probability and Statistics for Computer Science

Cited by 34 publications

References 0 publications

SDRSAC: Semidefinite-Based Randomized Approach for Robust Point Cloud Registration Without Correspondences

SDRSAC: Semidefinite-Based Randomized Approach for Robust Point Cloud Registration Without Correspondences

Dynamic core-periphery structure of information sharing networks in entorhinal cortex and hippocampus

Deep Learning for Stock Market Trading: A Superior Trading Strategy?

Contact Info

Product

Resources

About