Filters








94 Hits in 2.1 sec

Deep Argumentative Explanations [article]

Emanuele Albini, Piyawat Lertvittayakumjorn, Antonio Rago, Francesca Toni
2021 arXiv   pre-print
We refer to our novel explanations collectively as Deep Argumentative eXplanations (DAXs in short), given that they reflect the deep structure of the underlying NNs and that they are defined in terms of  ...  Despite the recent, widespread focus on eXplainable AI (XAI), explanations computed by XAI methods tend to provide little insight into the functioning of Neural Networks (NNs).  ...  We thank Kristijonas Cyras for helpful comments on earlier versions of this paper.  ... 
arXiv:2012.05766v4 fatcat:66qelbm4wzenzbfxllqckxswza

Compositional Generalization via Neural-Symbolic Stack Machines [article]

Xinyun Chen, Chen Liang, Adams Wei Yu, Dawn Song, Denny Zhou
2020 arXiv   pre-print
It contains a neural network to generate traces, which are then executed by a symbolic stack machine enhanced with sequence manipulation operations.  ...  To tackle this issue, we propose the Neural-Symbolic Stack Machine (NeSS).  ...  Related Work There has been an increasing interest in studying the compositional generalization of deep neural networks for natural language understanding [26, 24, 4, 41] .  ... 
arXiv:2008.06662v2 fatcat:5hcytak7zjdj5eotga3yzwgpyi

Forecasting of future stock prices using neural networks and genetic algorithms

Stelios A. Mitilineos, Panayiotis G. Artikis
2017 International Journal of Decision Sciences, Risk and Management  
Furthermore, we evaluate the use of modified GAs as a stand-alone tool for prediction, but also the use of GAs as neural network training and optimising tools.  ...  Based on a large body of work that is present in the literature, we develop, test and present a set of neural networks for predicting future stock market index values.  ...  neural networks with backpropagation training (see online version for colours) RMSE CoD  ... 
doi:10.1504/ijdsrm.2017.084002 fatcat:t27ltswvdnavjkjxjqhyekbzle

Seven challenges for harmonizing explainability requirements [article]

Jiahao Chen, Victor Storchan
2021 arXiv   pre-print
Regulators have signalled an interest in adopting explainable AI(XAI) techniques to handle the diverse needs for model governance, operational servicing, and compliance in the financial services industry  ...  in XAI and argue that based on our current understanding of the field, the use of XAI techniques in practice necessitate a highly contextualized approach considering the specific needs of stakeholders for  ...  DAX: Deep Argumentative eXplanation for Neural Networks. (2020). arXiv:2012.05766 http://arxiv.org/abs/2012.05766 [5] David Alvarez-Melis and Tommi S. Jaakkola. 2018.  ... 
arXiv:2108.05390v1 fatcat:2umqli4eknh7xlflheddl5nttu

Application of Artificial Intelligence in Stock Market Forecasting: A Critique, Review, and Research Agenda

Ritika Chopra, Gagan Deep Sharma
2021 Journal of Risk and Financial Management  
We conclude by establishing a research agenda for potential financial market analysts, artificial intelligence, and soft computing scholarship.  ...  This paper reviews 148 studies utilizing neural and hybrid-neuro techniques to predict stock markets, categorized based on 43 auto-coded themes obtained using NVivo 12 software.  ...  Acknowledgments: We would like to extend our gratitude towards Guru Gobind Singh Indraprastha University for providing us with the facilities and the infrastructure.  ... 
doi:10.3390/jrfm14110526 fatcat:jthkt5lwnfa3vmh43vtgoo3kue

Reproducibility and a unifying explanation: Lessons from the shape bias

Sarah C. Kucker, Larissa K. Samuelson, Lynn K. Perry, Hanako Yoshida, Eliana Colunga, Megan G. Lorenz, Linda B. Smith
2019 Infant Behavior and Development  
insight to underlying mechanisms, and that working to incorporate data from multiple labs is an important way to reveal how task variation and a child's individual pathway creates behavior-a key issue for  ...  To understand development, we need to go both deep and wide.  ...  an allencompassing explanation that remains coherent.  ... 
doi:10.1016/j.infbeh.2018.09.011 pmid:30343894 pmcid:PMC6393169 fatcat:i56owvulazdpjefcamfwy2f35a

What is the type-1/type-2 distinction?

Nick Chater
1997 Behavioral and Brain Sciences  
Overall, this distinction does not appear useful for machine learning or cognitive science.  ...  We are told that neural networks, or animats with neural network drivers, both having restricted classes of neural net architecture, cannot solve certain types of problems.  ...  A hidden assumption in C&T's target article that must be addressed is that what goes for neural networks goes for biological cognitive systems.  ... 
doi:10.1017/s0140525x97240021 fatcat:iuqgltj34nevpok4mtytazk6pm

Shedding Light on Black Box Machine Learning Algorithms: Development of an Axiomatic Framework to Assess the Quality of Methods that Explain Individual Predictions [article]

Milo Honegger
2018 arXiv   pre-print
With the proliferation of these explanation methods, it is, however, often unclear, which explanation method offers a higher explanation quality, or is generally better-suited for the situation at hand  ...  The main reason for this is that these methods boast remarkable predictive capabilities.  ...  We can e.g. have an explanation method for boosting (XGB), another one for bagging (random forests) and yet another one for deep neural networks (MLP).  ... 
arXiv:1808.05054v1 fatcat:vokg4heebbhajbw63aundexotq

Bootstrapping language acquisition

Omri Abend, Tom Kwiatkowski, Nathaniel J. Smith, Sharon Goldwater, Mark Steedman
2017 Cognition  
The learner thus demonstrates how statistical learning over structured representations can provide a unified account for these seemingly disparate phenomena.  ...  We thank Julia Hockenmaier and Luke Zettlemoyer for guidance in the early stages of this research, and Inbal Arnon, Jennifer Culbertson, and Ida Szubert for their feedback on a draft of this article.  ...  Many proponents of this approach invoke Artificial Neural Network (ANN) computational models as an explanation for how this could be done-see Elman et al. (1996) for examples-while others in both cognitive  ... 
doi:10.1016/j.cognition.2017.02.009 pmid:28412593 fatcat:3k3o4o2hw5cuddvqp7pfxzxsgy

From Discovering To Better Understanding The Relationship Between Brain And Behaviour

Eleftheria Dede, Ioannis Zalonis, Stylianos Gatzonis, Damianos Sakas
2017 Integrative Neuroscience Research  
networks.  ...  for future research.  ...  Ultimately, neuroimaging techniques provided the key to accurate correlations between cognitive processes and neural networks.  ... 
doi:10.35841/neuroscience.1.1.5-16 fatcat:epiheomzibhc5h2lgbub5vzjhm

Trading spaces: computation, representation, and the limits of uninformed learning

A Clark, C Thornton
1997 Behavioral and Brain Sciences  
Such mechanisms provide general (not task-specific) strategies for managing problems of type-2 complexity. Several such mechanisms are investigated.  ...  We are told that neural networks, or animats with neural network drivers, both having restricted classes of neural net architecture, cannot solve certain types of problems.  ...  A hidden assumption in C&T's target article that must be addressed is that what goes for neural networks goes for biological cognitive systems.  ... 
pmid:10096995 fatcat:weiwzeub4jhibmt7luvxde5lmm

Trading spaces: Computation, representation, and the limits of uninformed learning

Andy Clark, Chris Thornton
1997 Behavioral and Brain Sciences  
Such mechanisms provide general (not task-specific) strategies for managing problems of type-2 complexity. Several such mechanisms are investigated.  ...  We are told that neural networks, or animats with neural network drivers, both having restricted classes of neural net architecture, cannot solve certain types of problems.  ...  A hidden assumption in C&T's target article that must be addressed is that what goes for neural networks goes for biological cognitive systems.  ... 
doi:10.1017/s0140525x97000022 fatcat:irxrqvftjbg5xcxhgd6wp2i3rm

Précis of Semantic Cognition: A Parallel Distributed Processing Approach

Timothy T. Rogers, James L. McClelland
2008 Behavioral and Brain Sciences  
The explanations our theory offers for these phenomena are illustrated with reference to a simple feedforward connectionist model.  ...  Second, deep insights at the neural level will only be possible once we have a deep understanding of the computations supported by structured semantic representations.  ...  Hadley [2000] for relationships between classical computational models and neural networks; and Sarle [1994] for an overview of equivalencies between neural networks and different regression techniques  ... 
doi:10.1017/s0140525x0800589x fatcat:stacla5uy5hzddeclg24ocnkn4

Feature inference and the causal structure of categories

Bob Rehder, Russell C. Burnett
2005 Cognitive Psychology  
Our experiments were designed to test proposals that causal knowledge is represented psychologically as Bayesian networks.  ...  (e.g., see Cartwright, 1993, for arguments related to this idea; and Hausman & Woodward, 1999 , for a response).  ...  Fig. 8 . 8 Alternative causal models for the common cause network of Fig. 1. (A) Feature Uncertainity Model for Common Cause network. (B) Underlying Mechanism Model for Common Cause network.  ... 
doi:10.1016/j.cogpsych.2004.09.002 pmid:15826612 fatcat:6bsdtslr6bbhnli3iza4gehkpy

Tail-risk protection: Machine Learning meets modern Econometrics [article]

Bruno Spilak, Wolfgang Karl Härdle
2021 arXiv   pre-print
In this paper, we present a dynamic tail risk protection strategy that targets a maximum predefined level of risk measured by Value-At-Risk while controlling for participation in bull market regimes.  ...  Neural networks We first use two different neural network architectures for the trading signals prediction, the Multi-Layer Perceptron (MLP) and the Long short-term memory neural network (LSTM) architecture  ...  In particular, LSTM neural networks are state-of-the-art for many applications such as speech recognition, text extraction, translation or handwriting recognition, since plain recurrent neural networks  ... 
arXiv:2010.03315v4 fatcat:6zruv4nwx5d6jo3nvrgi6nohlu
« Previous Showing results 1 — 15 out of 94 results