Artificial Intelligence Review

Stock market fluctuations have a great impact on various economic and financial activities worldwide. Accurate prediction of stock prices plays a decisive role in constructing the investment decision or risk hedging. However, accurate prediction of the stock price is a thorny task, because stock price fluctuations are non-linear and chaotic. In order to promote the accuracy of stock price prediction, a multi-factor two-stage deep learning integrated prediction system based on intelligent optimization and feature clustering is proposed to predict stock price in this paper. Firstly, a multi-factor analysis is carried out to select a variety of factors that have an impact on the stock price, and adopt the extreme gradient boosting (XGBoost) algorithm to eliminate factors with low correlation. The second step is to apply the idea of classification prediction to cluster the filtered feature set. Further, multiple parameters of long short-term memory (LSTM) are optimized by genetic algorithm (GA), and multiple GA-LSTM models are obtained by training each clustering result. Finally, the results of each class predicted by the GA-LSTM model are nonlinearly integrated to acquire the final prediction model, which is applied to the prediction of the test set. The experimental results indicate that the performance of the proposed model outperforms other baseline models in China's two stock markets and the New York stock exchange. At the same time, these results fully prove that the prediction model proposed by us possesses more reliable and better predictive ability.

Humour is a valid subject for research in artificial intelligence, as it is one of the more complex of human behaviours. Although philosophers and others have discussed humour for centuries, it is only very recently that computational work has begun in this field, so the state of the art is still rather basic. Much of the research has concentrated on humour expressed verbally, and there has been some emphasis on models based on “incongruity”. Actual implementations have involved puns of very limited forms. It is not clear that computerised jokes could enhance user interfaces in the near future, but there is a role for computer modelling in testing symbolic accounts of the structure of humorous texts. A major problem is the need for a humour-processing program to have knowledge of the world, and reasoning abilities.

Researchers are defining new types of interactions between humans and machine learning algorithms generically called human-in-the-loop machine learning. Depending on who is in control of the learning process, we can identify: active learning, in which the system remains in control; interactive machine learning, in which there is a closer interaction between users and learning systems; and machine teaching, where human domain experts have control over the learning process. Aside from control, humans can also be involved in the learning process in other ways. In curriculum learning human domain experts try to impose some structure on the examples presented to improve the learning; in explainable AI the focus is on the ability of the model to explain to humans why a given solution was chosen. This collaboration between AI models and humans should not be limited only to the learning process; if we go further, we can see other terms that arise such as Usable and Useful AI. In this paper we review the state of the art of the techniques involved in the new forms of relationship between humans and ML algorithms. Our contribution is not merely listing the different approaches, but to provide definitions clarifying confusing, varied and sometimes contradictory terms; to elucidate and determine the boundaries between the different methods; and to correlate all the techniques searching for the connections and influences between them.

Individual learning in an environment where more than one agent exist is a chal-lengingtask. In this paper, a single learning agent situated in an environment where multipleagents exist is modeled based on reinforcement learning. The environment is non-stationaryand partially accessible from an agents' point of view. Therefore, learning activities of anagent is influenced by actions of other cooperative or competitive agents in the environment.A prey-hunter capture game that has the above characteristics is defined and experimentedto simulate the learning process of individual agents. Experimental results show that thereare no strict rules for reinforcement learning. We suggest two new methods to improve theperformance of agents. These methods decrease the number of states while keeping as muchstate as necessary.

Linear kernel Support Vector Machine Recursive Feature Elimination (SVM-RFE) is known as an excellent feature selection algorithm. Nonlinear SVM is a black box classifier for which we do not know the mapping function $${\Phi}$$ explicitly. Thus, the weight vector w cannot be explicitly computed. In this paper, we proposed a feature selection algorithm utilizing Support Vector Machine with RBF kernel based on Recursive Feature Elimination(SVM-RBF-RFE), which expands nonlinear RBF kernel into its Maclaurin series, and then the weight vector w is computed from the series according to the contribution made to classification hyperplane by each feature. Using $${w_i^2}$$ as ranking criterion, SVM-RBF-RFE starts with all the features, and eliminates one feature with the least squared weight at each step until all the features are ranked. We use SVM and KNN classifiers to evaluate nested subsets of features selected by SVM-RBF-RFE. Experimental results based on 3 UCI and 3 microarray datasets show SVM-RBF-RFE generally performs better than information gain and SVM-RFE.

Neural networks have recently been proposed for the construction of navigation interfaces for Information Retrieval systems. In this paper, we give an overview of some current research in this area. Most of the cited approaches use (variants) of the well-known Kohonen network. The Kohonen network implements a topology-preserving dimensionality-reducing mapping, which can be applied for information visualization. We identify a number of problems in the application of Kohonen networks for Information Retrieval, most notably scalability, reliability and retrieval effectiveness. To solve these problems we propose to use the Growing Cell Structures network, a variant of the Kohonen network which shows a more flexible adaptation to the domain structure. This network was tested on two standard test-collections, using a combined recall and precision measure, and compared to traditional IR methods such as the Vector Space Model and various clustering algorithms. The network performs at a competitive level of effectiveness, and is suitable for visualization purposes. However, the incremental training procedures for the networks result in a reliability problem, and the approach is computationally intensive. Also, the utility of the resulting maps for navigation will need further improvement.

This paper surveys research in developing computational models for integrating linguistic and visual information. It begins with a discussion of systems which have been actually implemented and continues with computationally motivated theories of human cognition. Since existing research spans several disciplines (e.g., natural language understanding, computer vision, knowledge representation), as well as several application areas, an important contribution of this paper is to categorize existing research based on inputs and objectives. Finally, some key issues related to integrating information from two such diverse sources are outlined and related to existing research. Throughout, the key issue addressed is the correspondence problem, namely how to associate visual events with words and vice versa.

Recent findings in neuroscience suggest an overlap between those brain regions involved in the control and execution of movement and those activated during the perception of another’s movement. This so called ‘mirror neuron’ system is thought to underlie our ability to automatically infer the goals and intentions of others by observing their actions. Kilner et al. (Curr Biol 13(6):522–525, 2003) provide evidence for a human ‘mirror neuron’ system by showing that the execution of simple arm movements is affected by the simultaneous perception of another’s movement. Specifically, observation of ‘incongruent’ movements made by another human, but not by a robotic arm, leads to greater variability in the movement trajectory than observation of movements in the same direction. In this study we ask which aspects of the observed motion are crucial to this interference effect by comparing the efficacy of real human movement to that of sparse ‘point-light displays’. Eight participants performed whole arm movements in both horizontal and vertical directions while observing either the experimenter or a virtual ‘point-light’ figure making arm movements in the same or in a different direction. Our results, however, failed to show an effect of ‘congruency’ of the observed movement on movement variability, regardless of whether a human actor or point-light figure was observed. The findings are discussed, and future directions for studies of perception-action coupling are considered.

In recent years, the fruit industry has become an important part of agricultural development, and fruit harvesting is a key stage in the production process. However, picking fruits during the harvest season is always a major challenge. In order to solve the challenges of time-consuming, costly, and inefficient fruit picking, researchers have conducted a lot of studies on automatic fruit picking equipment. Existing picking technologies still require further research and development to improve efficiency and reduce fruit damage. Aiming at the efficient and non-destructive picking of fruits, this paper reviews machine vision and mechanical fruit picking technology and the current research status, including the current application status, equipment structure, working principle, picking process, and experimental results. As a promising tool, machine vision technology has been widely researched and applied due to its low hardware cost and rich visual information. With the development of science and technology, automated fruit picking technology integrates information technology, integrates automatic perception, transmission, control, and operation, etc., saves manpower costs, and continuously promotes the development of modern agriculture in the direction of refinement of equipment technology, automation, and intelligence. Finally, the challenges faced by automated fruit picking are discussed, and future development is looked forward to with a view to contributing to its sustainable development.