N. Shayanfar; V. Derhami; M. Rezaeian
Abstract
In video prediction it is expected to predict next frame of video by providing a sequence of input frames. Whereas numerous studies exist that tackle frame prediction, suitable performance is not still achieved and therefore the application is an open problem. In this article multiscale processing is ...
Read More
In video prediction it is expected to predict next frame of video by providing a sequence of input frames. Whereas numerous studies exist that tackle frame prediction, suitable performance is not still achieved and therefore the application is an open problem. In this article multiscale processing is studied for video prediction and a new network architecture for multiscale processing is presented. This architecture is in the broad family of autoencoders. It is comprised of an encoder and decoder. A pretrained VGG is used as an encoder that processes a pyramid of input frames at multiple scales simultaneously. The decoder is based on 3D convolutional neurons. The presented architecture is studied by using three different datasets with varying degree of difficulty. In addition, the proposed approach is compared to two conventional autoencoders. It is observed that by using the pretrained network and multiscale processing results in a performant approach.
H. Khodadadi; V. Derhami
Abstract
A prominent weakness of dynamic programming methods is that they perform operations throughout the entire set of states in a Markov decision process in every updating phase. This paper proposes a novel chaos-based method to solve the problem. For this purpose, a chaotic system is first initialized, and ...
Read More
A prominent weakness of dynamic programming methods is that they perform operations throughout the entire set of states in a Markov decision process in every updating phase. This paper proposes a novel chaos-based method to solve the problem. For this purpose, a chaotic system is first initialized, and the resultant numbers are mapped onto the environment states through initial processing. In each traverse of the policy iteration method, policy evaluation is performed only once, and only a few states are updated. These states are proposed by the chaos system. In this method, the policy evaluation and improvement cycle lasts until an optimal policy is formulated in the environment. The same procedure is performed in the value iteration method, and only the values of a few states proposed by the chaos are updated in each traverse, whereas the values of other states are left unchanged. Unlike the conventional methods, an optimal solution can be obtained in the proposed method by only updating a limited number of states which are properly distributed all over the environment by chaos. The test results indicate the improved speed and efficiency of chaotic dynamic programming methods in obtaining the optimal solution in different grid environments.
H.6.2.2. Fuzzy set
N. Mohammadkarimi; V. Derhami
Abstract
This paper proposes fuzzy modeling using obtained data. Fuzzy system is known as knowledge-based or rule-bases system. The most important part of fuzzy system is rule-base. One of problems of generation of fuzzy rule with training data is inconsistence data. Existence of inconsistence and uncertain states ...
Read More
This paper proposes fuzzy modeling using obtained data. Fuzzy system is known as knowledge-based or rule-bases system. The most important part of fuzzy system is rule-base. One of problems of generation of fuzzy rule with training data is inconsistence data. Existence of inconsistence and uncertain states in training data causes high error in modeling. Here, Probability fuzzy system presents to improvement the above challenge. A zero order Sugeno fuzzy model used as fuzzy system structure. At first by using clustering obtains the number of rules and input membership functions. A set of candidate amounts for consequence parts of fuzzy rules is considered. Considering each pair of training data, according which rules fires and what is the output in the pair, the amount of probability of consequences candidates are change. In the next step, eligibility probability of each consequence candidate for all rules is determined. Finally, using these obtained probability, two probable outputs is generate for each input. The experimental results show superiority of the proposed approach rather than some available well-known approaches that makes reduce the number of rule and reduce system complexity.
H.6.3.2. Feature evaluation and selection
A. Zangooei; V. Derhami; F. Jamshidi
Abstract
Phishing is one of the luring techniques used to exploit personal information. A phishing webpage detection system (PWDS) extracts features to determine whether it is a phishing webpage or not. Selecting appropriate features improves the performance of PWDS. Performance criteria are detection accuracy ...
Read More
Phishing is one of the luring techniques used to exploit personal information. A phishing webpage detection system (PWDS) extracts features to determine whether it is a phishing webpage or not. Selecting appropriate features improves the performance of PWDS. Performance criteria are detection accuracy and system response time. The major time consumed by PWDS arises from feature extraction that is considered as feature cost in this paper. Here, two novel features are proposed. They use semantic similarity measure to determine the relationship between the content and the URL of a page. Since suggested features don't apply third-party services such as search engines result, the features extraction time decreases dramatically. Login form pre-filer is utilized to reduce unnecessary calculations and false positive rate. In this paper, a cost-based feature selection is presented as the most effective feature. The selected features are employed in the suggested PWDS. Extreme learning machine algorithm is used to classify webpages. The experimental results demonstrate that suggested PWDS achieves high accuracy of 97.6% and short average detection time of 120.07 milliseconds.
H.3.12. Distributed Artificial Intelligence
M. Rezaei; V. Derhami
Abstract
Nonnegative Matrix Factorization (NMF) algorithms have been utilized in a wide range of real applications. NMF is done by several researchers to its part based representation property especially in the facial expression recognition problem. It decomposes a face image into its essential parts (e.g. nose, ...
Read More
Nonnegative Matrix Factorization (NMF) algorithms have been utilized in a wide range of real applications. NMF is done by several researchers to its part based representation property especially in the facial expression recognition problem. It decomposes a face image into its essential parts (e.g. nose, lips, etc.) but in all previous attempts, it is neglected that all features achieved by NMF do not need for recognition problem. For example, some facial parts do not have any useful information regarding the facial expression recognition. Addressing this challenge of defining and calculating the contributions of each part, the Shapley value is used. It is applied for identifying the contribution of each feature in the classification problem; then, affects less features are removed. Experiments on the JAFFE dataset and MUG Facial Expression Database as benchmarks of facial expression datasets demonstrate the effectiveness of our approach.
H.3.15.3. Evolutionary computing and genetic algorithms
M. B. Dowlatshahi; V. Derhami
Abstract
A combinatorial auction is an auction where the bidders have the choice to bid on bundles of items. The WDP in combinatorial auctions is the problem of finding winning bids that maximize the auctioneer’s revenue under the constraint that each item can be allocated to at most one bidder. The WDP ...
Read More
A combinatorial auction is an auction where the bidders have the choice to bid on bundles of items. The WDP in combinatorial auctions is the problem of finding winning bids that maximize the auctioneer’s revenue under the constraint that each item can be allocated to at most one bidder. The WDP is known as an NP-hard problem with practical applications like electronic commerce, production management, games theory, and resources allocation in multi-agent systems. This has motivated the quest for efficient approximate algorithms both in terms of solution quality and computational time. This paper proposes a hybrid Ant Colony Optimization with a novel Multi-Neighborhood Local Search (ACO-MNLS) algorithm for solving Winner Determination Problem (WDP) in combinatorial auctions. Our proposed MNLS algorithm uses the fact that using various neighborhoods in local search can generate different local optima for WDP and that the global optima of WDP is a local optima for a given its neighborhood. Therefore, proposed MNLS algorithm simultaneously explores a set of three different neighborhoods to get different local optima and to escape from local optima. The comparisons between ACO-MNLS, Genetic Algorithm (GA), Memetic Algorithm (MA), Stochastic Local Search (SLS), and Tabu Search (TS) on various benchmark problems confirm the efficiency of ACO-MNLS in the terms of solution quality and computational time.
G.4. Information Storage and Retrieval
V. Derhami; J. Paksima; H. Khajah
Abstract
The main challenge of a search engine is ranking web documents to provide the best response to a user`s query. Despite the huge number of the extracted results for user`s query, only a small number of the first results are examined by users; therefore, the insertion of the related results in the first ...
Read More
The main challenge of a search engine is ranking web documents to provide the best response to a user`s query. Despite the huge number of the extracted results for user`s query, only a small number of the first results are examined by users; therefore, the insertion of the related results in the first ranks is of great importance. In this paper, a ranking algorithm based on the reinforcement learning and user`s feedback called RL3F are considered. In the proposed algorithm, the ranking system has been considered to be the agent of learning system and selecting documents to display to the user is as the agents’ action. The reinforcement signal in the system is calculated according to a user`s clicks on documents. Action-value values of the proposed algorithm are computed for each feature. In each learning cycle, the documents are sorted out for the next query, and according to the document in the ranked list, documents are selected at random to show the user. Learning process continues until the training is completed. LETOR3 benchmark is used to evaluate the proposed method. Evaluation results indicated that the proposed method is more effective than other methods mentioned for comparison in this paper. The superiority of the proposed algorithm is using several features of document and user`s feedback simultaneously.
Mohaddeseh Dashti; Vali Derhami; Esfandiar Ekhtiyari
Abstract
Yarn tenacity is one of the most important properties in yarn production. This paper addresses modeling of yarn tenacity as well as optimally determining the amounts of the effective inputs to produce yarn with desired tenacity. The artificial neural network is used as a suitable structure for tenacity ...
Read More
Yarn tenacity is one of the most important properties in yarn production. This paper addresses modeling of yarn tenacity as well as optimally determining the amounts of the effective inputs to produce yarn with desired tenacity. The artificial neural network is used as a suitable structure for tenacity modeling of cotton yarn with 30 Ne. As the first step for modeling, the empirical data is collected for cotton yarns. Then, the structure of the neural network is determined and its parameters are adjusted by back propagation method. The efficiency and accuracy of the neural model is measured based on percentage of error as well as coefficient determination. The obtained experimental results show that the neural model could predicate the tenacity with less than 3.5% error. Afterwards, utilizing genetic algorithms, a new method is proposed for optimal determination of input values in yarn production to reach the desired tenacity. We conducted several experiments for different ranges with various production cost functions. The proposed approach could find the best input values to reach the desired tenacity considering the production costs.
Mohammad AllamehAmiri; Vali Derhami; Mohammad Ghasemzadeh
Abstract
Quality of service (QoS) is an important issue in the design and management of web service composition. QoS in web services consists of various non-functional factors, such as execution cost, execution time, availability, successful execution rate, and security. In recent years, the number of available ...
Read More
Quality of service (QoS) is an important issue in the design and management of web service composition. QoS in web services consists of various non-functional factors, such as execution cost, execution time, availability, successful execution rate, and security. In recent years, the number of available web services has proliferated, and then offered the same services increasingly. The same web services are distinguished based on their quality parameters. Also, clients usually demand more value added services rather than those offered by single, isolated web services. Therefore, selecting a composition plan of web services among numerous plans satisfies client requirements and has become a challenging and time-consuming problem. This paper has proposed a new composition plan optimizer with constraints based on genetic algorithm. The proposed method can find the composition plan that satisfies user constraints efficiently. The performance of the method is evaluated in a simulated environment.
