Explore the vast range of titles available.

Background Biomedical semantic indexing is important for information retrieval and many other research fields in bioinformatics. It annotates biomedical citations with Medical Subject Headings. In face of unbalanced category distribution in the training data, sampling methods are difficult to apply for semantic indexing task. Results In this paper, we present a novel deep serial multi-task learning model. The primary task treats the biomedical semantic indexing as a multi-label text classification issue that considers the relations of the labels. The auxiliary task is a regression task that predicts the MeSH number of the citation and provides hints for the network to make it converge faster. The experimental results on the BioASQ-Task5A open dataset show that our model outperforms the state-of-the-art solution “MTI”, proposed by the US National Library of Medicine. Further, it not only achieves the highest precision among all the solutions in BioASQ-Task5A but also has faster convergence speed compared with some naive deep learning methods. Conclusions Rather than parallel in an ordinary multi-task structure, the tasks in our model are serial and tightly coupled. It can achieve satisfied performance without any handcrafted feature.

New observations are made about two lower bound schemes for single-machine min-sum scheduling problems. We find that the strongest bound of those provided by transportation problem relaxations can be computed by solving a linear program. We show the equivalence of this strongest bound and the bound provided by the LP relaxation of the time-indexed integer programming formulation. These observations lead to a new lower bound scheme that yields fast approximation of the time-indexed bound. Several techniques are developed to facilitate the effective use of the new lower bound in branch-and-bound. Numerical experiments are conducted on 375 benchmark problems of the total weighted tardiness problem from OR-Library. Results obtained with our new method are spectacular; we are able to solve all 125 open problems to optimality. [PUBLICATION ABSTRACT]

Linear programming (LP)-based relaxations have proven to be useful in enumerative solution procedures for NP -hard min–sum scheduling problems. We take a dual viewpoint of the time-indexed integer linear programming (ILP) formulation for these problems. Previously proposed Lagrangian relaxation methods and a time decomposition method are interpreted and synthesized under this view. Our new results aim to find optimal or near-optimal dual solutions to the LP relaxation of the time-indexed formulation, as recent advancements made in solving this ILP problem indicate the utility of dual information. Specifically, we develop a procedure to compute optimal dual solutions using the solution information from Dantzig–Wolfe decomposition and column generation methods, whose solutions are generally nonbasic. As a byproduct, we also obtain, in some sense, a crossover method that produces optimal basic primal solutions. Furthermore, the dual view naturally leads us to propose a new polynomial-sized relaxation that is applicable to both integer and real-valued problems. The obtained dual solutions are incorporated in branch-and-bound for solving the total weighted tardiness scheduling problem, and their efficacy is evaluated and compared through computational experiments involving test problems from OR-Library.

In view of the fact that there are disadvantages in that the class number must be determined in advance, the value of learning rates are hard to fix, etc., when using traditional competitive neural networks (CNNs) in electronic noses (E-noses), an optimized CNN method was presented. The optimized CNN was established on the basis of the optimum class number of samples according to the changes of the Davies and Bouldin (DB) value and it could increase, divide, or delete neurons in order to adjust the number of neurons automatically. Moreover, the learning rate changes according to the variety of training times of each sample. The traditional CNN and the optimized CNN were applied to five kinds of sorted vinegars with an E-nose. The results showed that optimized network structures could adjust the number of clusters dynamically and resulted in good classifications.

A central problem in the field of robotics is to develop real-time planning and control algorithms for autonomous systems to behave intelligently under uncertainty. While classical optimal control provides a general theoretical framework, it relies on strong assumption of full knowledge of the system dynamics and environments. Alternatively, modern reinforcement learning (RL) offers a computational framework for controlling autonomous systems with minimal prior knowledge and user intervention. However, typical RL approaches require many interactions with the physical systems, and suffer from slow convergence. Furthermore, both optimal control and RL have the difficulty of scaling to high-dimensional state and action spaces.In order to address these challenges, we present probabilistic trajectory optimization methods for solving optimal control problems for systems with unknown or partially known dynamics. Our methods share two key characteristics: (1) we incorporate explicit uncertainty into modeling, prediction and decision making using Gaussian processes; (2) our algorithms bypass the curse of dimensionality via local approximation of the value function or linearization of the Hamilton-Jacobi-Bellman (HJB) equation. Compared to related approaches, our methods offer superior combination of data efficiency and scalability. We present experimental results and comparative analyses to demonstrate the strengths of the proposed methods.In addition, we develop fast Bayesian approximate inference methods which enable probabilistic trajectory optimizer to perform real-time receding horizon control. It can be used to train deep neural network controllers that map raw observations to actions directly. We show that our approach can be used to perform high-speed off-road autonomous driving with low-cost sensors, and without on-the-fly planning and optimization.

In this paper, we introduce a new job shop model that minimizes a well-motivated inventory measure while assuring on-time job deliveries. For this new problem, we introduce precise notation and formalization. A decomposition scheme is discussed in detail, which is subsequently utilized in a new shifting bottleneck procedure (SBP) for the problem. In addition to SBP, we propose another heuristic method based on successive insertion of operations. Algorithms are fine tuned through experimentation. Moreover, the two heuristic procedures are compared in terms of computation time and solution quality, using disguised actual factory data.

Remote terminal unit (RTU) is the core device of the monitor system in hydrology and water resources. Different devices often have different communication protocols in the application layer, which results in the difficulty in information analysis and communication networking. Therefore, we introduced the idea of software defined hardware, and abstracted the common feature of mainstream communication protocols of RTU application layer, and proposed a uniformed common protocol model. Then, various communication protocol algorithms of application layer are modularized according to the model. The executable codes of these algorithms are labeled by the virtual functions and stored in the flash chips of embedded CPU to form the protocol stack. According to the configuration commands to initialize the RTU communication systems, it is able to achieve dynamic assembling and loading of various application layer communication protocols of RTU and complete the efficient transport of sensor data from RTU to central station when the data acquisition protocol of sensors and various external communication terminals remain unchanged.

We present two nonparametric approaches to Kullback-Leibler (KL) control, or linearly-solvable Markov decision problem (LMDP) based on Gaussian processes (GP) and Nystr\"om approximation. Compared to recently developed parametric methods, the proposed data-driven frameworks feature accurate function approximation and efficient on-line operations. Theoretically, we derive the mathematical connection of KL control based on dynamic programming with earlier work in control theory which relies on information theoretic dualities for the infinite time horizon case. Algorithmically, we give explicit optimal control policies in nonparametric forms, and propose on-line update schemes with budgeted computational costs. Numerical results demonstrate the effectiveness and usefulness of the proposed frameworks.

In a globally competitive economy, controlling production costs and maximizing responsiveness are of critical importance to the extended enterprise supply chain. A key objective for suppliers is to sustain the lowest possible work-in-process and finished goods inventories while simultaneously ensuring the prompt deliveries of parts and components to the downstream original equipment manufacturer (OEM). Achieving this objective requires optimized production scheduling that utilizes integrated demand information from the OEM to model and optimize the continuous status of the supplier's labor, equipment, and parts. This research was motivated by a complex problem that arose during the course of an industry project with the supplier development team of John Deere Horicon Works (the world's largest manufacturer of premium ride-on lawn tractors located in Horicon, Wisconsin) to help one of their suppliers revamp its scheduling function. We formulate the problem using a new performance measure called the total inventory. To solve the problem, we proceed along two directions. First, we seek heuristic solutions to a multi-machine model that takes into account various factors including alternate machines, shifts and down times, and labor required to run machines and to perform machine setups. Second, we consider simplified models that admit theoretical analyses and exact solution methods. Specifically, two heuristic algorithms are introduced, based on insertion and shifting bottleneck, respectively. We also make the connection between our new models and existing ones, and investigate relevant computational complexity issues. Furthermore, we develop a new lower bound and a branch-and-bound algorithm for solving a particular single-machine subproblem that is key to our multi-machine models as well as a class of existing multi-machine models. Our results can be employed to develop more effective heuristic methods as well as enumerative procedures for such multi-machine problems.

Trajectory optimization of a controlled dynamical system is an essential part of autonomy, however many trajectory optimization techniques are limited by the fidelity of the underlying parametric model. In the field of robotics, a lack of model knowledge can be overcome with machine learning techniques, utilizing measurements to build a dynamical model from the data. This paper aims to take the middle ground between these two approaches by introducing a semi-parametric representation of the underlying system dynamics. Our goal is to leverage the considerable information contained in a traditional physics based model and combine it with a data-driven, non-parametric regression technique known as a Gaussian Process. Integrating this semi-parametric model with model predictive pseudospectral control, we demonstrate this technique on both a cart pole and quadrotor simulation with unmodeled damping and parametric error. In order to manage parametric uncertainty, we introduce an algorithm that utilizes Sparse Spectrum Gaussian Processes (SSGP) for online learning after each rollout. We implement this online learning technique on a cart pole and quadrator, then demonstrate the use of online learning and obstacle avoidance for the dubin vehicle dynamics.