Universal Face Recognition Using Multiple Deep Learning Agent and Lazy Learning Algorithm

Mainstream face recognition systems have a problem regarding the disparity of recognizing faces from different races and ethnic backgrounds. This problem is caused by the imbalances in the proportion of racial representations found in mainstream datasets. Hence, the research proposes using a multi-agent system to overcome this problem. The system employs several face recognition agents according to the number of races that are necessary to make data encodings for the classification process. The first step in implementing this system is to develop a race classifier. The number of races is arbitrary or determined differently in a caseby-case scenario. The race classifier determines which face recognition agent will try to recognize the face in the query. Each face recognition agent is trained using a different dataset according to their assigned race, so they have different parts in the system. The research utilizes lazy learning algorithms as the final classifier to accommodate a system with the constant data flow of the database. The experiment divides the data into three racial groups, which are black, Asian, and white. The experiment concludes that dividing face recognition tasks based on racial groups into several face recognition models has better performance than a single model with the same dataset with the same imbalances in racial representation. The multiple agent system achieves 85% on the Face Recognition Rate (FRR), while the single pipeline model achieves only 80.83% using the same dataset.

Monolingual and Cross-Lingual Intent Detection without Training Data in Target Languages

Due to recent DNN advancements, many NLP problems can be effectively solved using transformer-based models and supervised data. Unfortunately, such data is not available in some languages. This research is based on assumptions that (1) training data can be obtained by the machine translating it from another language; (2) there are cross-lingual solutions that work without the training data in the target language. Consequently, in this research, we use the English dataset and solve the intent detection problem for five target languages (German, French, Lithuanian, Latvian, and Portuguese). When seeking the most accurate solutions, we investigate BERT-based word and sentence transformers together with eager learning classifiers (CNN, BERT fine-tuning, FFNN) and lazy learning approach (Cosine similarity as the memory-based method). We offer and evaluate several strategies to overcome the data scarcity problem with machine translation, cross-lingual models, and a combination of the previous two. The experimental investigation revealed the robustness of sentence transformers under various cross-lingual conditions. The accuracy equal to ~0.842 is achieved with the English dataset with completely monolingual models is considered our top-line. However, cross-lingual approaches demonstrate similar accuracy levels reaching ~0.831, ~0.829, ~0.853, ~0.831, and ~0.813 on German, French, Lithuanian, Latvian, and Portuguese languages.

Rich and lazy learning of task representations in brains and neural networks

How do neural populations code for multiple, potentially conflicting tasks? Here, we used computational simulations involving neural networks to define “lazy” and “rich” coding solutions to this multitasking problem, which trade off learning speed for robustness. During lazy learning the input dimensionality is expanded by random projections to the network hidden layer, whereas in rich learning hidden units acquire structured representations that privilege relevant over irrelevant features. For context-dependent decision-making, one rich solution is to project task representations onto low-dimensional and orthogonal manifolds. Using behavioural testing and neuroimaging in humans, and analysis of neural signals from macaque prefrontal cortex, we report evidence for neural coding patterns in biological brains whose dimensionality and neural geometry are consistent with the rich learning regime.

3D car shape reconstruction from a contour sketch using GAN and lazy learning

Abstract3D car models are heavily used in computer games, visual effects, and even automotive designs. As a result, producing such models with minimal labour costs is increasingly more important. To tackle the challenge, we propose a novel system to reconstruct a 3D car using a single sketch image. The system learns from a synthetic database of 3D car models and their corresponding 2D contour sketches and segmentation masks, allowing effective training with minimal data collection cost. The core of the system is a machine learning pipeline that combines the use of a generative adversarial network (GAN) and lazy learning. GAN, being a deep learning method, is capable of modelling complicated data distributions, enabling the effective modelling of a large variety of cars. Its major weakness is that as a global method, modelling the fine details in the local region is challenging. Lazy learning works well to preserve local features by generating a local subspace with relevant data samples. We demonstrate that the combined use of GAN and lazy learning produces is able to produce high-quality results, in which different types of cars with complicated local features can be generated effectively with a single sketch. Our method outperforms existing ones using other machine learning structures such as the variational autoencoder.

Decision branch joint venture ex-fold T-z re-validation

Comparing classifier performances may seem a banal affair but makes a side show in machine learning. Usually the paired t-test is used. It requires that two classifiers were run simultaneously or this was simulated. This is not always possible and then entails creating a superstructure only for that purpose. However, the utility of t-test in the given context is altogether doubted. The literature on alternatives is much involved. This does not measure up to the scale of the issue. In this paper the topics in connection with accuracy calculation are surveyed once more, emphasizing the result variation. The known technique of multifold cross-validation is exemplified. A simplified methodology for comparison of classifier performances is proposed. It is based on the accuracy mean and variance and calculating differences between objects defined in these terms. It is being applied to the naive Bayesian and decision tree classifiers implemented on different platforms. The lazy learning approach, applicable to decision trees in discrete domains, is closely followed with an imposition of how it can be improved. Examples are given from the field of health diagnostics.

Lazy Learning Associative Classification with WkNN and DWkNN Algorithm

One of the algorithms, which prudently denote better outcomes than the traditional associative classification systems, is the Lazy learning associative classification (LLAC), where the processing of training data is delayed until a test instance is received, whereas in eager learning, before receiving queries, the system begins to process training data. Traditional method assumes that all items within a transaction is same, which is not always true. This paper recommends a new framework called lazy learning associative classification with WkNN (LLAC_WkNN) which uses weighted kNN method with LLAC, that gives a subset of rules when LLAC is applied to the dataset. In order to predict the class label of the unseen test case, the weighted kNN (WkNN) algorithm is then applied to this generated subset. This creates the enhanced accuracy of the classifier. The WkNN also gives an outlier more weight. By applying Dual Distance Weight to LLAC named as LLAC_DWkNN, this limitation of WkNN is resolved. LLAC_DWkNN gives less weightage to outliers, which improve the accuracy of the classifier, further. This algorithm has been applied to different datasets and the experiment results demonstrate that the proposed method is efficient as compared to the traditional and other existing systems.