This paper tries to resolve long waiting time to find a matching person in player versus player mode of online sports games, such as baseball, soccer and basketball. In player versus player mode, game playing AI which is instead of player needs to be not just smart as human but also show variety to improve user experience against AI. Therefore a need to design game playing AI agents with diverse personalized styles rises. To this end, we propose a personalized game AI which encodes user style vectors and card style vectors with a general DNN, named UCSM-DNN. Extensive experiments show that UCSM-DNN shows improved performance in terms of personalized styles, which enrich user experiences. UCSM-DNN has already been integrated into popular mobile baseball game: MaguMagu 2021 as personalized game AI.

In this paper, we propose on-device voice command assistants for mobile games to increase user experiences even in hands-busy situations such as driving and cooking. Since most of the current mobile games cost large memory (e.g. more than 1GB memory), so it is necessary to reduce memory usage further to integrate voice commands systems on mobile clients. Therefore a need to design an on-device automatic speech recognition system that costs minimal memory and CPU resources rises. To this end, we apply cross layer parameter sharing to Conformer, named MONICA2 which results in lower memory usage for on-device speech recognition. MONICA2 reduces the number of parameters of deep neural network by 58%, with minimal recognition accuracy degradation measured in word error rate on Librispeech benchmark. As an on-device voice command user interface, MONICA2 costs only 12.8MB mobile memory and the average inference time for 3-seconds voice command is about 30ms, which is profiled in Samsung Galaxy S9. As far as we know, MONICA2 is the most memory efficient yet accurate on-device speech recognition which could be applied to various applications such as mobile games, IoT devices, etc.

Multi-task learning is a prominent methodology for learning relationship of multiple related tasks to improve the overall generalization performance. In literature, there are several works of applying nonnegative matrix factorization (NMF) to multi-task clustering (MTC) for its useful behavior as a clustering method. When NMF is applied to a data matrix, the matrix is decomposed into a product of two factor matrices, where one corresponds to basis matrix (cluster centers) and the other is associated with encoding matrix (cluster indicator variables). In this paper, we consider clustering activities of daily living (ADL) as a single task. Given ADL data recorded from different people under the same conditions, our motivation is that related tasks share many common action patterns, while each task also has its own unique action characteristics. Our goal is to learn cross-task basis matrices that capture shared information across multiple related tasks, while try to find within-task basis matrices encoding task-dependent characteristics, which will not appear in any other tasks. To this end, we propose a method for multi-task nonnegative matrix factorization (MTNMF), which enforces cross-task bases to be as close as possible, on the contrary, keeps within-task bases as far as possible. We develop efficient multiplicative update rules to learn MTNMF and numerical experiments on ADL dataset indicate that MTNMF improves clustering performance in multi-task environment.

This paper presents a fast and robust plane feature extraction and matching technique for RGB-D type sensors. We propose three algorithm components required to utilize the plane features in an online Simultaneous Localization and Mapping (SLAM) problem: fast plane extraction, frame-to-frame constraint estimation, and plane merging. For the fast plane extraction, we estimate local surface normals and curvatures by a simple spherical model and then segment points using a modified flood fill algorithm. In plane parameter estimation, we suggest a new uncertainty estimation method which is robust against the measurement bias, and also introduce a fast boundary modeling method. We associate the plane features based on both the parameters and the spatial coverage, and estimate the stable constraints by the cost function with a regulation term. Also, our plane merging technique provides a way of generating local maps that are useful for estimating loop closure constraints. We have performed real-world experiments at our lab environment. The results demonstrate the efficiency and robustness of the proposed algorithm.

Kernel nonnegative matrix factorization (KNMF) is a recent kernel extension of NMF, where matrix factorization is carried out in a reproducing kernel Hilbert space (RKHS) with a feature mapping φ(·). Given a data matrix X ∈ ℝm×n, KNMF seeks a decomposition, φ(X) ≈ UVT, where the basis matrix takes the form U = φ(X)W and parameters W ∈ ℝn×rand V ∈ ℝ+n×rare estimated without explicit knowledge of φ(·). As in most of kernel methods, the performance of KNMF also heavily depends on the choice of kernel. In order to alleviate the kernel selection problem when a single kernel is used, we present multiple kernel NMF (MKNMF) where two learning problems are jointly solved in unsupervised manner: (1) learning the best convex combination of kernel matrices; (2) learning parameters W and V. We formulate multiple kernel learning in MKNMF as a linear programming and estimate W and V using multiplicative updates as in KNMF. Experiments on benchmark face datasets confirm the high performance of MKNMF over several existing variants of NMF, in the task of feature extrac tion for face classification.