Sanghee Lee

Yuri Kim, Yewon Choi, Yujung Gil et al.

Although the many efforts to apply deep reinforcement learning to query optimization in recent years, there remains room for improvement as query optimizers are complex entities that require hand-designed tuning of workloads and datasets. Recent research present learned query optimizations results mostly in bulks of single workloads which focus on picking up the unique traits of the specific workload. This proves to be problematic in scenarios where the different characteristics of multiple workloads and datasets are to be mixed and learned together. Henceforth, in this paper, we propose BitE, a novel ensemble learning model using database statistics and metadata to tune a learned query optimizer for enhancing performance. On the way, we introduce multiple revisions to solve several challenges: we extend the search space for the optimal Abstract SQL Plan(represented as a JSON object called ASP) by expanding hintsets, we steer the model away from the default plans that may be biased by configuring the experience with all unique plans of queries, and we deviate from the traditional loss functions and choose an alternative method to cope with underestimation and overestimation of reward. Our model achieves 19.6% more improved queries and 15.8% less regressed queries compared to the existing traditional methods whilst using a comparable level of resources.

Heemoon Yoon, Sang-Hee Lee, Mira Park

TensorFlow is an open-source framework for deep learning dataflow and contains application programming interfaces (APIs) of voice analysis, natural language process, and computer vision. Especially, TensorFlow object detection API in computer vision field has been widely applied to technologies of agriculture, engineering, and medicine but barriers to entry of the framework usage is still high through command-line interface (CLI) and code for amateurs and beginners of information technology (IT) field. Therefore, this is aim to develop an user friendly Graphical Framework for object detection API on TensorFlow which is called TensorFlow Graphical Framework (TF-GraF). The TF-GraF provides independent virtual environments according to user accounts in server-side, additionally, execution of data preprocessing, training, and evaluation without CLI in client-side. Furthermore, hyperparameter setting, real-time observation of training process, object visualization of test images, and metrics evaluations of test data can also be operated via TF-GraF. Especially, TF-GraF supports flexible model selection of SSD, Faster-RCNN, RFCN, and Mask-RCNN including convolutional neural networks (inceptions and ResNets) through GUI environment. Consequently, TF-GraF allows anyone, even without any previous knowledge of deep learning frameworks, to design, train and deploy machine intelligence models without coding. Since TF-GraF takes care of setting and configuration, it allows anyone to use deep learning technology for their project without spending time to install complex software and environment.

Tae Jun Ham, Sung Jun Jung, Seonghak Kim et al.

With the increasing computational demands of neural networks, many hardware accelerators for the neural networks have been proposed. Such existing neural network accelerators often focus on popular neural network types such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs); however, not much attention has been paid to attention mechanisms, an emerging neural network primitive that enables neural networks to retrieve most relevant information from a knowledge-base, external memory, or past states. The attention mechanism is widely adopted by many state-of-the-art neural networks for computer vision, natural language processing, and machine translation, and accounts for a large portion of total execution time. We observe today's practice of implementing this mechanism using matrix-vector multiplication is suboptimal as the attention mechanism is semantically a content-based search where a large portion of computations ends up not being used. Based on this observation, we design and architect A3, which accelerates attention mechanisms in neural networks with algorithmic approximation and hardware specialization. Our proposed accelerator achieves multiple orders of magnitude improvement in energy efficiency (performance/watt) as well as substantial speedup over the state-of-the-art conventional hardware.