Deep learning techniques have revolutionised artificial systems’ performance on myriad tasks, from playing Go to medical diagnosis. Recent developments have extended such successes to natural language processing, an area once deemed beyond such systems’ reach. Despite their different goals, these successes have suggested that such systems may be pertinent to theoretical linguistics. The competence/performance distinction presents a fundamental barrier to such inferences. While DL systems are trained on linguistic performance, linguistic theories are aimed at competence. Such a barrier (...) has traditionally been sidestepped by assuming a fairly close correspondence: performance as competence plus noise. I argue this assumption is unmotivated. Competence and performance can differ arbitrarily. Thus, we should not expect DL models to illuminate linguistic theory. (shrink)

Deep learning (DL) is a statistical technique for pattern classification through which AI researchers train artificial neural networks containing multiple layers that process massive amounts of data. I present a three-level account of explanation that can be reasonably expected from DL models in cognitive neuroscience and that illustrates the explanatory dynamics within a future-biased research program (Feest Philosophy of Science 84:1165–1176, 2017 ; Doerig et al. Nature Reviews: Neuroscience 24:431–450, 2023 ). By relying on the mechanistic framework (Craver (...) Explaining the brain: Mechanisms and the mosaic unity of neuroscience. Clarendon, 2007 ; Stinson Eppuor si muove: Doing history and philosophy of Science with Peter Machamer. Springer, 2017, The Routledge Handbook of the computational mind. Routledge, 2018 ), I develop an account that corresponds to the stages of mechanism discovery, i.e., our shifting epistemic position and epistemic goals, and propose a representative model for each level. Generic, theoretical DL models at _Level 1_ address the general features of a cognitive phenomenon through exploration and provide us with how-possibly explanations. On the other hand, DL models at _Level 2_ either identify the interaction between the features or represent the epistemic stage when the researcher is still unsure if the modeled features are crucial or arbitrary. These models should provide us with how-plausibly explanations. Finally, specific DL models of specific brain areas, i.e., ersatz models filled with relevant cognitive and neuroscientific details, are at _Level 3_. At this level, a researcher can advance how-actually explanations of cognitive phenomena. The main strength of this account is that it elucidates both _global explanatory dynamics_ and _local explanatory dynamics_ (cf. Dresow Synthese 199:10441–10474, 2021 ). The former occurs when the transition between levels happens in accordance with the process of obtaining more details about a particular cognitive phenomenon through multiple DL models. The latter, meanwhile, involves cases in which the transition between levels takes place within a single DL model by elucidating internal mechanisms (e.g., using Explainable AI techniques for rendering models more transparent). (shrink)

Brain–computer interfaces (BCIs) utilizing machine learning techniques are an emerging technology that enables a communication pathway between a user and an external system, such as a computer. Owing to its practicality, electroencephalography (EEG) is one of the most widely used measurements for BCI. However, EEG has complex patterns and EEG-based BCIs mostly involve a cost/time-consuming calibration phase; thus, acquiring sufficient EEG data is rarely possible. Recently, deep learning (DL) has had a theoretical/practical impact on BCI research because (...) of its use in learning representations of complex patterns inherent in EEG. Moreover, algorithmic advances in DL facilitate short/zero-calibration in BCI, thereby suppressing the data acquisition phase. Those advancements include data augmentation (DA), increasing the number of training samples without acquiring additional data, and transfer learning (TL), taking advantage of representative knowledge obtained from one dataset to address the so-called data insufficiency problem in other datasets. In this study, we review DL-based short/zero-calibration methods for BCI. Further, we elaborate methodological/algorithmic trends, highlight intriguing approaches in the literature, and discuss directions for further research. In particular, we search forgenerative model-based andgeometric manipulation-based DA methods. Additionally, we categorize TL techniques in DL-based BCIs intoexplicitandimplicitmethods. Our systematization reveals advances in the DA and TL methods. Among the studies reviewed herein, ~45% of DA studies used generative model-based techniques, whereas ~45% of TL studies used explicit knowledge transferring strategy. Moreover, based on our literature review, we recommend an appropriate DA strategy for DL-based BCIs and discuss trends of TLs used in DL-based BCIs. (shrink)

Processing data streams arriving at high speed requires the development of models that can provide fast and accurate predictions. Although deep neural networks are the state-of-the-art for many machine learning tasks, their performance in real-time data streaming scenarios is a research area that has not yet been fully addressed. Nevertheless, much effort has been put into the adaption of complex deep learning (DL) models to streaming tasks by reducing the processing time. The design of the asynchronous (...) dual-pipeline DL framework allows making predictions of incoming instances and updating the model simultaneously, using two separate layers. The aim of this work is to assess the performance of different types of DL architectures for data streaming classification using this framework. We evaluate models such as multi-layer perceptrons, recurrent, convolutional and temporal convolutional neural networks over several time series datasets that are simulated as streams at different speeds. In addition, we evaluate how the different architectures react to concept drifts typically found in evolving data streams. The obtained results indicate that convolutional architectures achieve a higher performance in terms of accuracy and efficiency, but are also the most sensitive to concept drifts. (shrink)

As the development of artificial intelligence technology, the deep-learning -based Virtual Reality technology, and DL technology are applied in human-computer interaction, and their impacts on modern film and TV works production and audience psychology are analyzed. In film and TV production, audiences have a higher demand for the verisimilitude and immersion of the works, especially in film production. Based on this, a 2D image recognition system for human body motions and a 3D recognition system for human body motions (...) based on the convolutional neural network algorithm of DL are proposed, and an analysis framework is established. The proposed systems are simulated on practical and professional datasets, respectively. The results show that the algorithm's computing performance in 2D image recognition is 7–9 times higher than that of the Open Pose method. It runs at 44.3 ms in 3D motion recognition, significantly lower than the Open Pose method's 794.5 and 138.7 ms. Although the detection accuracy has dropped by 2.4%, it is more efficient and convenient without limitations of scenarios in practical applications. The AI-based VR and DL enriches and expands the role and application of computer graphics in film and TV production using HCI technology theoretically and practically. (shrink)

Forecasting blood glucose values for patients can help prevent hypoglycemia and hyperglycemia events in advance. To this end, this study proposes an intelligent ensemble deep learning system to predict BG values in 15, 30, and 60 min prediction horizons based on historical BG values collected via continuous glucose monitoring devices as an endogenous factor and carbohydrate intake and insulin administration information as exogenous factors. Although there are numerous deep learning algorithms available, this study applied five algorithms, (...) namely, recurrent neural network, which is optimized for sequence data, and RNN-based algorithms, stacked LSTM, bidirectional LSTM, and gated recurrent unit). Then, a genetic algorithm was applied to the five prediction models to optimize their weights through ensemble techniques and to yield the final predicted BG values. The performance of the proposed model was compared to that of the autoregressive integrated moving average model as a baseline. The results show that the proposed model significantly outperforms the baseline in terms of the root mean square error and continuous glucose error grid analysis. For the valid 29 diabetic patients for the multivariate models, the RMSE was 11.08, 19.25, and 31.30 mg/DL for 15, 30, and 60 min PH, respectively. When the same data were applied to univariate models, the RMSE was 11.28, 19.99, and 33.13 mg/DL for 15, 30, and 60 min PH, respectively. Both the univariate and multivariate models showed a statistically significant difference compared with the baseline at a 5% statistical significance level. Instead of using a model with a single algorithm, applying a GA based on each output of a model with multiple algorithms was found to play a significant role in improving model performance. (shrink)

With the rapid expansion of Internet technology, this research aims to explore the teaching strategies of ceramic art for contemporary students. Based on deep learning, an automatic question answering system is established, new teaching strategies are analyzed, and the Internet is combined with the automatic QA system to help students solve problems encountered in the process of learning. Firstly, the related theories of DL and personalized learning are analyzed. Among DL-related theories, Back Propagation Neural Network, Convolutional (...) Neural Network, Long Short-Term Memory, and Gated Recurrent Unit are compared to implement a single model and a mixed model. Secondly, the collected student questions are selected and processed, and experimental parameters in different models are set for comparative experiments. Experiments reveal that the average accuracy and Mean Reciprocal Rank of traditional retrieval methods can only reach about 0.5. In the basic neural network, the average accuracy of LSTM and GRU structural models is about 0.81, which can achieve better results. Finally, the accuracy of the hybrid model can reach about 0.82, and the accuracy and MRR of the Bidirectional Gated Recurrent Unit Network-Attention model are 0.87 and 0.89, respectively, achieving the best results. The established DL model meets the requirements of the online automatic QA system, improves the teaching system, and helps students better understand and solve problems in the ceramic art courses. (shrink)

The purpose is to explore new cultivation modes of college students’ national cognition and cultural acceptance. Deep learning technology and Educational Psychology theory are introduced, and the influence of art journal reading on college students’ national cognition and cultural acceptance is analyzed under Educational Psychology. Firstly, the background of Educational Psychology, national cognition and cultural acceptance, and learning system are discussed following a literature review. The DL technology is introduced to construct the journal reading guidance system. The (...) system can provide users with art journals and record the user habits like reading duration and preferences. Secondly, hypotheses are proposed, and a questionnaire survey is designed, with 12 specific indicators to investigate and collect research data. Finally, the collected data are analyzed. The results show that women’s cognition of Chinese traditional culture, Chinese excellent revolutionary culture, and Chinese national identity is higher than that of men. By comparison, men’s cognition of Chinese advanced socialist culture is higher than women’s. After using the journal reading guidance system, the cognition of female college students on traditional Chinese culture is improved by 16.3%. Before and after reading art journals, the overall national cognition and cultural acceptance of Minority students are higher than that of Han students. The overall cognition of Literature and History students is higher than that of Science and Engineering students in traditional Chinese culture and China’s excellent revolutionary culture and lower in advanced Chinese socialist culture and Chinese national identity. The overall cognition of college students’ party members to the advanced socialist culture is higher than league members. As students read more art journals through the guidance system, their overall national cognition and cultural acceptance have increased. Therefore, reading art journals can promote college students’ national cognition and cultural acceptance. A national cognition and cultural acceptance promotion system that conforms to the current situation of college students is constructed. The finding provides a reference for developing complex emotion recognition technology in human-computer interaction. (shrink)

This work aims to reform legal teaching in Colleges and Universities and improve law students’ comprehensive quality. In the context of Educational Psychology research, Deep Learning theory is integrated into legal instructional design. Following a theoretical review of EPSY and DL, the current situation and problems of college legal teaching are understood based on the Law School in a University in Shanghai through auditing, communication, and investigation methods. The theoretical research results are integrated into the ID. The teaching (...) content is divided into language information module, wisdom skills module, cognitive module, action skills module, and attitude module. Each module is divided into three teaching methods, and all teaching methods are combined into the proposed legal ID. Finally, the proposed legal ID is applied in the legal classroom of the Law School in a University in Shanghai. Overall, seventy students are recruited and grouped into Class A and Class B. Class A uses the proposed legal ID, and Class B does not. The scores of Classes A and B are compared. After a semester, the average score of Class A has increased from 68 to 71.11 points. The covariance has decreased from 61.66 to 51.42. When the confidence level is set to 0.95, the confidence interval of class A has increased from 65.26–70.74 to 68.62–73.61. By comparison, the average score of Class B dropped from 68.14 to 68.11 points. The covariance has decreased from 60.24 to 41.76. When the confidence level is set to 0.95, the confidence interval of class B has changed from 65.44–70.85 to 65.86–70.37, without significant improvement. Therefore, the proposed legal ID based on DL theory is scientific and effective. This work has certain reference significance for optimizing the ID of CAUs and improving the comprehensive quality of college-student talents. (shrink)

This paper sheds light on the shift that is taking place from the practice of ‘coding’, namely developing programs as conventional in the software community, to the practice of ‘curing’, an activity that has emerged in the last few years in Deep Learning (DL) and that amounts to curing the data regime to which a DL model is exposed during training. Initially, the curing paradigm is illustrated by means of a study-case on autonomous vehicles. Subsequently, the shift from (...) coding to curing is analysed taking into consideration the epistemological notions, central in the philosophy of computer science, of function, implementation, and correctness. First, it is illustrated how, in the curing paradigm, the functions performed by the trained model depend much more on dataset curation rather than on the model algorithms which, in contrast with the coding paradigm, do not comply with requested specifications. Second, it is highlighted how DL models cannot be considered implementations according to any of the available definitions of implementation that follow an intentional theory of functions. Finally, it is argued that DL models cannot be evaluated in terms of their correctness but rather in their experimental computational validity. (shrink)

Environmental landscaping is known to build, plan, and manage landscapes that consider the ecology of a site and produce gardens that benefit both people and the rest of the ecosystem. Landscaping and the environment are combined in landscape design planning to provide holistic answers to complex issues. Seeding native species and eradicating alien species are just a few ways humans influence the region’s ecosystem. Landscape architecture is the design of landscapes, urban areas, or gardens and their modification. It comprises the (...) construction of urban and rural landscapes via coordinating the creation and management of open spaces and economics, finding a job, and working within a confined project budget. There was a lot of discussion about global warming and water shortages. There is a lot of hope to be found even in the face of seemingly insurmountable obstacles. AI is becoming more significant in many urban landscape planning and design elements with the advent of web 4.0 and Human-Centred computing. It created a virtual reality-based landscape to create deep neural networks (DNNs) to make deep learning (DL) more user-friendly and efficient. Users may only manipulate physical items in this environment to manually construct neural networks. These setups are automatically converted into a model, and the real-time testing set is reported and aware of the DNN models that users are producing. This research presents a novel strategy for combining DL-DNN with landscape architecture, providing a long-term solution to the problem of environmental pollution. Carbon dioxide levels are constantly checked when green plants are in and around the house. Plants, on either hand, remove toxins from the air, making it easier to maintain a healthy environment. Human-centered Artificial Intelligence-based web 4.0 may be used to assess and evaluate the data model. The study findings can be sent back into the design process for further modification and optimization. (shrink)

Deep learning and artificial intelligence are jointly applied to concrete piano teaching for children to comprehensively promote modern piano teaching and improve the overall teaching quality. First, the teaching environment and the functions of the intelligent piano are expounded. Then, a piano note onset detection method is proposed based on the convolution neural network. The network can analyze the time-frequency of the input piano music signal by transforming the original time-domain waveform of the piano music signal into the (...) frequency distribution varying with time. Besides, it can detect the note onset at a stable state after 8 × 104 iterations. Moreover, an intelligent piano teaching method is designed to teach Jingle Bells for 40 preschool children aged 4–6 years. Finally, a questionnaire survey is performed to investigate the teaching situation, including the learning interest and learning effect of children and learning feedback from parents. The results show that 80% of children like smart music scores, 82% of children like intelligent piano lessons with games, and 84% of children can learn actively in the intelligent piano class. Besides, 85% of parents believe that their children are more interested in learning piano. In general, the intelligent piano teaching method effectively combines DL with AI to realize the overall optimization of piano performance. It is widely favored by preschool children and their parents and plays an important role in improving the interest of preschool children in piano learning. (shrink)

This work intends to solve the problem that the traditional education system cannot reasonably adjust the educational integration of children with the arrival of labor force in a short time, and support the education of migrant children in the education policy to integrate them into the local educational environment as soon as possible. Firstly, this work defines the surplus labor force and MC. Secondly, the principles of Artificial Intelligence and Deep Learning are introduced. Thirdly, it analyzes the education (...) of MC and relevant policies, and the data of the education effect of MC are collected and the evaluation effect model is built. Finally, the evaluation model of MC’s education effect is applied to test the effect of EP. The results show that using AI technology combined with DL technology to model the education effect of MC can establish an effective and accurate evaluation model of the education effect of MC, effectively evaluate the impact of local education policies on the education of MC, and give an effective effect analysis of relevant education policies in each period. The result of Adaptive Resonance Theory –Back Propagation algorithm is 65 ∼ 96%, which is much higher than the efficiency of traditional algorithms. This shows that the education integration evaluation model of MC based on AI technology and DL technology can effectively and accurately evaluate the integration effect of MC on the local education system, and then provide reference for local and even national adjustment of education policies. The results provide a new idea for the application of new technology in EP. (shrink)

This paper investigates if and to what point it is possible to trade on news sentiment and if deep learning, given the current hype on the topic, would be a good tool to do so. DL is built explicitly for dealing with significant amounts of data and performing complex tasks where automatic learning is a necessity. Thanks to its promise to detect complex patterns in a dataset, it may be appealing to those investors that are looking to (...) improve their trading process. Moreover, DL and specifically LSTM seem a good pick from a linguistic perspective too, given its ability to “remember” previous words in a sentence. After having explained how DL models are built, we will use this tool for forecasting the market sentiment using news headlines. The prediction is based on the Dow Jones industrial average by analyzing 25 daily news headlines available between 2008 and 2016, which will then be extended up to 2020. The result will be the indicator used for developing an algorithmic trading strategy. The analysis will be performed on two specific cases that will be pursued over five time-steps and the testing will be developed in real-world scenarios. (shrink)

In order to improve the efficiency of teaching and learning in Colleges and Universities, this work combines the Browser/Server framework with Model View Presenter technology to build a college student–oriented micro-video teaching system based on Deep Learning and Human–Computer Interaction technology. Firstly, it makes an in-depth analysis of the problems in the classroom teaching of Chinese CAUs. Three functional modules are designed for the micro-video online teaching platform: video management, user learning, and system management. Then, it (...) uses MVP technology to analyze the use-cases of these three functional modules in detail. Based on this, the micro-video online teaching platform is designed using the B/S framework. The teaching platform interface layer realizes the HCI between the platform and users. The business logic layer responds to the user requests submitted and returns the processing results to the interface layer. Finally, the function test and stress test of each module of the micro-video online teaching platform is carried out. The test results show that the response time of the proposed micro-video teaching platform increases with the number of users. Under the peak concurrent users, the system response time is 6 s, without abnormalities. Meanwhile, the proposed teaching platform has improved students' satisfaction with classroom teaching by nearly 15% and improved the compactness of the college classroom by nearly 12%. When the number of virtual users increases and the number of services increases linearly, the Random Access Memory and Central Processing Unit growth rate is significantly lower than that of the number of services. These outcomes indicate that many system resources are reused, and the system has good scalability, which can meet users' needs for the network video teaching system. The proposed teaching platform provides a new idea for applying DL and HCI technology in researching college students' micro-video teaching mode. (shrink)

The increasing integration of Machine Learning (ML) techniques into clinical care, driven in particular by Deep Learning (DL) using Artificial Neural Nets (ANNs), promises to reshape medical practice on various levels and across multiple medical fields. Much recent literature examines the ethical consequences of employing ML within medical and psychiatric practice but the potential impact on psychiatric diagnostic systems has so far not been well-developed. In this article, we aim to explore the challenges that arise from the (...) recent use of ANNs for the old problems of psychiatric nosology. To enable an empirically supported critical reflection on the topic, we conducted semi-structured qualitative interviews with Swiss and German experts in computational psychiatry. Here, we report our findings structured around two themes, namely (1) the possibility of using ML for defining or refining of psychiatric classification, and (2) the desirability of employing ML for psychiatric nosology. We discuss these themes by relating them to recent debates about network theory for psychiatric nosology and show why empirical research in the field should critically reflect on its contribution to psychopathology research. In sum, we argue that beyond technical, regulatory, and ethical challenges, philosophical reflection is crucial to harness the potential of ML in psychiatry. (shrink)

The role of artificial intelligence in the Turing test is to imitate human beings to such an extent that people will not realize it is a machine. With the rise of deep learning (a subcategory of AI), the situation is changing rapidly as the new systems do not focus on imitating human intelligence but emphasize thorough solutions to specific issues. The main difference between predefined AI and deep learning (DL) is that these systems are self-learning (...) and have verifiable results. Firstly, we need to analyse the application of the Turing test in the Loebner Prize because, there, the primary emphasis is on aspects of human intelligence – learning, reasoning and understanding. Secondly, in the Turing test, only general intelligence is considered, and this can be questionable. If DL does not possess this form of intelligence, by this reasoning, we should consider it unintelligent. However, is such understanding correct? The third and last aspect questions whether the Turing test is beneficial for an AI designed for specific tasks because the results do not bring any new data and conclusions. (shrink)

The role of artificial intelligence in the Turing test is to imitate human beings to such an extent that people will not realize it is a machine. With the rise of deep learning (a subcategory of AI), the situation is changing rapidly as the new systems do not focus on imitating human intelligence but emphasize thorough solutions to specific issues. The main difference between predefined AI and deep learning (DL) is that these systems are self-learning (...) and have verifiable results. Firstly, we need to analyse the application of the Turing test in the Loebner Prize because, there, the primary emphasis is on aspects of human intelligence – learning, reasoning and understanding. Secondly, in the Turing test, only general intelligence is considered, and this can be questionable. If DL does not possess this form of intelligence, by this reasoning, we should consider it unintelligent. However, is such understanding correct? The third and last aspect questions whether the Turing test is beneficial for an AI designed for specific tasks because the results do not bring any new data and conclusions. (shrink)

Deep learning (DL) has become increasingly central to science, primarily due to its capacity to quickly, efficiently, and accurately predict and classify phenomena of scientific interest. This paper seeks to understand the principles that underwrite scientists’ epistemic entitlement to rely on DL in the first place and argues that these principles are philosophically novel. The question of this paper is not whether scientists can be justified in trusting in the reliability of DL. While today’s artificial intelligence exhibits characteristics (...) common to both scientific instruments and scientific experts, this paper argues that the familiar epistemic categories that justify belief in the reliability of instruments and experts are distinct, and that belief in the reliability of DL cannot be reduced to either. Understanding what can justify belief in AI reliability represents an occasion and opportunity for exciting, new philosophy of science. (shrink)

The development of machine learning and deep learning (DL) in the field of AI (artificial intelligence) is the direct result of the advancement of nano-electronics. Machine learning is a function that provides the system with the capacity to learn from data without being programmed explicitly. It is basically a mathematical and probabilistic model. DL is part of machine learning methods based on artificial neural networks, simply called neural networks (NNs), as they are inspired by the (...) biological NNs that constitute organic brains. Despite its similarity to biological organs such as human brains, major problems arise in trying to attribute moral responsibility to autonomic systems based on hardware including nano-electronic devices, which are sought to replace humans (moral agents) in the context of AI. It is suggested that the required emotional environment which enables actions according to reasons in humans is not witnessed at AI devices. Though AI technology raises an enticing resemblance to human actions, this resemblance is to be considered with a skeptical eye. It is because actions are associated with reasons while causes are connected to operations. Human agents are capable of acting upon their reasons, while AI devices are limited only to operations as they are conditioned by their programming, which is considered as an embodiment of some causes. As moral responsibility goes hand in hand with the capacity to act (according to reasons), attributing moral responsibility to AI devices is revealed to be but a misleading metaphor. (shrink)

Speech is a primary means of human communication and one of the most basic features of human conduct. Voice is an important part of its subsystems. A speech disorder is a condition that affects the ability of a person to speak normally, which occasionally results in voice impairment with psychological and emotional consequences. Early detection of voice problems is a crucial factor. Computer-based procedures are less costly and easier to administer for such purposes than traditional methods. This study highlights the (...) following issues: recent studies, methods of voice pathology detection, machine learning and deep learning (DL) methods used in data classification, main datasets utilized, and the role of Internet of things (IoT) systems employed in voice pathology diagnosis. Moreover, this study presents different applications, open challenges, and recommendations for future directions of IoT systems and artificial intelligence (AI) approaches in the voice pathology diagnosis. Finally, this study highlights some limitations of voice pathology datasets in comparison with the role of IoT in the healthcare sector, which shows the urgent need to provide efficient approaches and easy and ideal medical diagnostic procedures and treatments of disease identification for doctors and patients. This review covered voice pathology taxonomy, detection techniques, open challenges, limitations, and recommendations for future directions to provide a clear background for doctors and patients. Standard databases, including the Massachusetts Eye and Ear Infirmary, Saarbruecken Voice Database, and the Arabic Voice Pathology Database, were used in most articles reviewed in this article. The classes, features, and main purpose for voice pathology identification are also highlighted. This study focuses on the extraction of voice pathology features, especially speech analysis, extends feature vectors comprising static and dynamic features, and converts these extended feature vectors into solid vectors before passing them to the recognizer. (shrink)

With the rapid emergence of the technology of deep learning, it was successfully used in different fields such as the aquatic product. New opportunities in addition to challenges can be created according to this change for helping data processing in the smart fish farm. This study focuses on deep learning applications and how to support different activities in aquatic like identification of the fish, species classification, feeding decision, behavior analysis, estimation size, and prediction of water quality. (...) Power and performance of computing with the analyzed given data are applied in the proposed DL method within fish farming. Results of the proposed method show the significance of contributions in deep learning and how automatic features are extracted. Still, there is a big challenge of using deep learning in an era of artificial intelligence. Training of the proposed method used a large number of labeled images got from the Fish4Knowledge dataset. The proposed method based on suitable feature extracted from the fish achieved good results in terms of recognition rate and accuracy. (shrink)