Search

In the absence of user profile information, recommender systems have to only rely on current session information for recommendation. E-commerce sites may use transitions between interactions in each session to improve recommendation. This situation is known as the session-based recommendation. It can be challenging due to the limited information and the uncertain user behavior. Recurrent Neural Networks (RNN) have become the state-of-the-art models for session-based recommendation due to their ability to model long sequences. Although powerful, RNN-based models suffer from learning complex transition between the interactions. To mitigate it, Graph Neural Networks (GNN) have been proposed for session-based recommendation. However, different sequences of interactions may lead to the same outcome especially on E-commerce sites, hence non-sequential interactions between items of the current session may improve the performance of a recommender system. To learn both the sequential and non-sequential transition interactions between the items in the current session, we proposed a GNN based model named GRASER. Specifically, the proposed model first learns the non-sequential and then the sequential transition interactions between the items of the current session using GNN in an end-to-end manner. Extensive experiments were carried out on two datasets: Yoochoose from the RecSys Challenge 2015 and Diginetica from CIKM Cup 2016. The results showed that the proposed model outperforms the other state-of-the-art models by 11% and 10% on MRR@20 on Yoochoose and Diginetica datasets respectively.

We take a step towards addressing the under-representation of the African continent in NLP research by creating the first large publicly available high-quality dataset for named entity recognition (NER) in ten African languages, bringing together a variety of stakeholders. We detail characteristics of the languages to help researchers understand the challenges that these languages pose for NER. We analyze our datasets and conduct an extensive empirical evaluation of state-of-the-art methods across both supervised and transfer learning settings. We release the data, code, and models in order to inspire future research on African NLP., Comment: Accepted to TACL 2021, pre-MIT Press publication version

 Abstract—The leaky least-mean-square (LLMS) algorithm was first proposed to mitigate the drifting problem of the least- mean-square (LMS) algorithm. Though the LLMS algorithm solves this problem, its performance is similar to that of the LMS algorithm. In this paper, we propose an improved version of the LLMS algorithm that brings better performance to the LLMS algorithm and similarly solves the problem of drifting in the LMS algorithm. This better performance is achieved at a negligible increase in the computational complexity. The performance of the proposed algorithm is compared to that of the conventional LLMS algorithm in a system identification and a noise cancellation settings in additive white and correlated, Gaussian and impulsive, noise environments. Index Terms—Leaky least-mean-square, system identification, noise cancellation.

This paper proposed a facial expression recognition approach based on Gabor wavelet transform. Gabor wavelet filter is first used as pre-processing stage for extraction of the feature vector representation. Dimensionality of the feature vector is reduced using Principal Component Analysis (PCA) and Local binary pattern (LBP) algorithms. Experiments were carried out of using Japanese female facial expression (JAFFE) database. In all experiments conducted using JAFFE database, results obtained reveal that GW+LBP has outperformed other approaches in this paper with an average recognition rate of 90% under the same experimental setting.

In this paper, a new sparse adaptive filtering algorithm is proposed. The proposed algorithm introduces a log-sum penalty term into the cost function of a mixed norm leaky least-mean-square (NLLMS) algorithm. The cost function of the NLLMS algorithm is expressed in terms of sum of exponentials with a leakage factor. As a result of the log-sum penalty, the performance of the proposed algorithm is high in sparse system identification settings, especially, when the unknown system is highly sparse. The performance of the proposed algorithm is compared to those of the reweighted-zero-attracting LMS (RZA-LMS) and the p-norm variable step-size LMS (PNVSSLMS) algorithms in sparse system identification settings. The proposed algorithm shows superior performance compared to the aforementioned algorithms.