Hierarchical forecasting (HF) is needed in many situations in the supply chain (SC) because managers often need different levels of forecasts at different levels of SC to make a decision. Top-Down (TD), Bottom-Up (BU) and Optimal Combination (COM) are common HF models. These approaches are static and often ignore the dynamics of the series while disaggregating them. Consequently, they may fail to perform well if the investigated group of time series are subject to large changes such as during the periods of promotional sales. We address the HF problem of predicting real-world sales time series that are highly impacted by promotion. We use three machine learning (ML) models to capture sales variations over time. Artificial neural networks (ANN), extreme gradient boosting (XGboost), and support vector regression (SVR) algorithms are used to estimate the proportions of lower-level time series from the upper level. We perform an in-depth analysis of 61 groups of time series with different volatilities and show that ML models are competitive and outperform some well-established models in the literature.

Most of the existing methods are based on semi-definite programming ( SDP), which is generally time-consuming and degrades the scalability, especially confronting large-scale data. T o overcome this challenge, we propose a stochastic algorithm called SVRG-SBB, which has the following features: i) achieving good scalability via dropping positive semi-definite ( PSD) constraints as serving a fast algorithm, i.e., stochastic variance reduced gradient ( SVRG) method, and ii) adaptive learning via introducing a new, adaptive step size called the stabilized Barzilai-Borwein ( SBB) step size. Theoretically, under some natural assumptions, we show the O ( 1 T) rate of convergence to a stationary point of the proposed algorithm, where T is the number of total iterations. Under the further Polyak- Łojasiewicz assumption, we can show the global linear convergence (i.e., exponentially fast converging to a global optimum) of the proposed algorithm. Numerous simulations and real-world data experiments are conducted to show the effectiveness of the proposed algorithm by comparing with the state-of-the-art methods, notably, much lower computational cost with good prediction performance. Index Terms --Ordinal Embedding, SVRG, Non-Convex Optimization, Barzilai-Borwein (BB) Step Size, .null 1 I NTRODUCTION O RDINAL embedding aims to learn the representation of data as points in a low-dimensional embedded space. Here the "low-dimensional" means the embedding-K. Ma is with the School of Computer Science and T echnology, University of Chinese Academy of Sciences, Beijing 100049, China, and with the Artificial Intelligence Research Center, Peng Cheng Laboratory, Shenzhen 518055, China, and part of this work was performed when he was in the Key Laboratory of Information Security, Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China, and in the School of Cyber Security, University of Chinese Academy of Sciences, Beijing 100049, China. Email: make@ucas.ac.cn - J. Zeng is with the School of Computer Information Engineering, Jiangxi Normal University, Nanchang, Jiangxi 330022, China, and part of this work was performed when he was with the Department of Mathematics, Hong Kong University of Science and T echnology, Clear Water Bay, Kowloon, Hong Kong.

Machine learning models can be used for pattern recognition in medical data in order to improve patient outcomes, such as the prediction of in-hospital mortality. Deep learning models, in particular, require large amounts of data for model training. However, the data is often collected at different hospitals and sharing is restricted due to patient privacy concerns. In this paper, we aimed to demonstrate the potential of distributed training in achieving state-of-the-art performance while maintaining data privacy. Our results show that training the model in the federated learning framework leads to comparable performance to the traditional centralised setting. We also suggest several considerations for the success of such frameworks in future work.

Although deep neural networks generally have fixed network structures, the concept of dynamic mechanism has drawn more and more attention in recent years. Attention mechanisms compute input-dependent dynamic attention weights for aggregating a sequence of hidden states. Dynamic network configuration in convolutional neural networks (CNNs) selectively activates only part of the network at a time for different inputs. In this paper, we combine the two dynamic mechanisms for text classification tasks. Traditional attention mechanisms attend to the whole sequence of hidden states for an input sentence, while in most cases not all attention is needed especially for long sequences. We propose a novel method called Gated Attention Network (GA-Net) to dynamically select a subset of elements to attend to using an auxiliary network, and compute attention weights to aggregate the selected elements. It avoids a significant amount of unnecessary computation on unattended elements, and allows the model to pay attention to important parts of the sequence. Experiments in various datasets show that the proposed method achieves better performance compared with all baseline models with global or local attention while requiring less computation and achieving better interpretability. It is also promising to extend the idea to more complex attention-based models, such as transformers and seq-to-seq models.

Asynchronous Advantage Actor Critic (A3C) is an effective Reinforcement Learning (RL) algorithm for a wide range of tasks, such as Atari games and robot control. The agent learns policies and value function through trial-and-error interactions with the environment until converging to an optimal policy. Robustness and stability are critical in RL; however, neural network can be vulnerable to noise from unexpected sources and is not likely to withstand very slight disturbances. We note that agents generated from mild environment using A3C are not able to handle challenging environments. Learning from adversarial examples, we proposed an algorithm called Adversary Robust A3C (AR-A3C) to improve the agent's performance under noisy environments. In this algorithm, an adversarial agent is introduced to the learning process to make it more robust against adversarial disturbances, thereby making it more adaptive to noisy environments. Both simulations and real-world experiments are carried out to illustrate the stability of the proposed algorithm. The AR-A3C algorithm outperforms A3C in both clean and noisy environments.

Greenhouse environment is the key to influence crops production. However, it is difficult for classical control methods to give precise environment setpoints, such as temperature, humidity, light intensity and carbon dioxide concentration for greenhouse because it is uncertain nonlinear system. Therefore, an intelligent close loop control framework based on model embedded deep reinforcement learning (MEDRL) is designed for greenhouse environment control. Specifically, computer vision algorithms are used to recognize growing periods and sex of crops, followed by the crop growth models, which can be trained with different growing periods and sex. These model outputs combined with the cost factor provide the setpoints for greenhouse and feedback to the control system in real-time. The whole MEDRL system has capability to conduct optimization control precisely and conveniently, and costs will be greatly reduced compared with traditional greenhouse control approaches.

At high latitudes, many cities adopt a centralized heating system to improve the energy generation efficiency and to reduce pollution. In multi-tier systems, so-called district heating, there are a few efficient approaches for the flow rate control during the heating process. In this paper, we describe the theoretical methods to solve this problem by deep reinforcement learning and propose a cloud-based heating control system for implementation. A real-world case study shows the effectiveness and practicability of the proposed system controlled by humans, and the simulated experiments for deep reinforcement learning show about 1985.01 gigajoules of heat quantity and 42276.45 tons of water are saved per hour compared with manual control.

Understanding each other is the key to success in collaboration. For humans, attributing mental states to others, the theory of mind, provides the crucial advantage. We argue for formulating human--AI interaction as a multi-agent problem, endowing AI with a computational theory of mind to understand and anticipate the user. To differentiate the approach from previous work, we introduce a categorisation of user modelling approaches based on the level of agency learnt in the interaction. We describe our recent work in using nested multi-agent modelling to formulate user models for multi-armed bandit based interactive AI systems, including a proof-of-concept user study.

Conformance checking techniques let us find out to what degree a process model and real execution data correspond to each other. In recent years, alignments have proven extremely useful in calculating conformance statistics. Most techniques to compute alignments provide an exact solution. However, in many applications, it is enough to have an approximation of the conformance value. Specifically, for large event data, the computing time for alignments is considerably long using current techniques which makes them inapplicable in reality. Also, it is no longer feasible to use standard hardware for complex processes. Hence, we need techniques that enable us to obtain fast, and at the same time, accurate approximation of the conformance values. This paper proposes new approximation techniques to compute approximated conformance checking values close to exact solution values in a faster time. Those methods also provide upper and lower bounds for the approximated alignment value. Our experiments on real event data show that it is possible to improve the performance of conformance checking by using the proposed methods compared to using the state-of-the-art alignment approximation technique. Results show that in most of the cases, we provide tight bounds, accurate approximated alignment values, and similar deviation statistics.

Abstraction reasoning is a long-standing challenge in artificial intelligence. Recent studies suggest that many of the deep architectures that have triumphed over other domains failed to work well in abstract reasoning. In this paper, we first illustrate that one of the main challenges in such a reasoning task is the presence of distracting features, which requires the learning algorithm to leverage counterevidence and to reject any of the false hypotheses in order to learn the true patterns. We later show that carefully designed learning trajectory over different categories of training data can effectively boost learning performance by mitigating the impacts of distracting features. Inspired by this fact, we propose feature robust abstract reasoning (FRAR) model, which consists of a reinforcement learning based teacher network to determine the sequence of training and a student network for predictions. Experimental results demonstrated strong improvements over baseline algorithms and we are able to beat the state-of-the-art models by 18.7% in the RAVEN dataset and 13.3% in the PGM dataset.