Joint Program and Layout Transformations to Enable Convolutional Operators on Specialized Hardware Based on Constraint Programming

The success of Deep Artificial Neural Networks (DNNs) in many domains created a rich body of research concerned with hardware accelerators for compute-intensive DNN operators. However, implementing such operators efficiently with complex hardware intrinsics such as matrix multiply is a task not yet automated gracefully. Solving this task often requires joint program and data layout transformations. First solutions to this problem have been proposed, such as TVM, UNIT, or ISAMIR, which work on a loop-level representation of operators and specify data layout and possible program transformations before the embedding into the operator is performed. This top-down approach creates a tension between exploration range and search space complexity, especially when also exploring data layout transformations such as im2col, channel packing, or padding. In this work, we propose a new approach to this problem. We created a bottom-up method that allows the joint transformation of both computation and data layout based on the found embedding. By formulating the embedding as a constraint satisfaction problem over the scalar dataflow, every possible embedding solution is contained in the search space. Adding additional constraints and optimization targets to the solver generates the subset of preferable solutions. An evaluation using the VTA hardware accelerator with the Baidu DeepBench inference benchmark shows that our approach can automatically generate code competitive to reference implementations. Further, we show that dynamically determining the data layout based on intrinsic and workload is beneficial for hardware utilization and performance. In cases where the reference implementation has low hardware utilization due to its fixed deployment strategy, we achieve a geomean speedup of up to × 2.813, while individual operators can improve as much as × 170.

Deformation Analysis of an Ultra-High Arch Dam under Different Water Level Conditions Based on Optimized Dynamic Panel Clustering

During the operation period, the deformation of an ultra-high arch dam is affected by the large fluctuation of the reservoir water level. Under the dual coupling of the ultra-high dam and the complex water level conditions, the traditional variational analysis method cannot be sufficiently applied to its deformation analysis. The deformation analysis of the ultra-high arch dam, however, is very important in order to judge the dam safety state. To analyze the deformation law of different parts of an ultra-high arch dam, the panel data clustering theory is used to construct a Spatio-temporal characteristic model of dam deformation. In order to solve the difficult problem of the fluctuating displacement of dam deformation with water level effect, three displacement component indexes (absolute quantity, growing, and fluctuation) are proposed to characterize dam deformation. To further optimize the panel clustering deformation model, the objective weight coefficient of clustering comprehensive distance is calculated based on the CRITIC (CRiteria Importance Through Inter-criteria Correlation) method. The zoning rules of the ultra-high arch dam are established by using the idea of the CSP (Constraint Satisfaction Problem) index, and the complex water level of the reservoir is simulated in the whole process. Finally, the dynamic cluster analysis of dam deformation is realized. Through a case study, three typical working conditions including the rapid rise and fall of water level and the normal operation are calculated, and the deformation laws of different deformation zones are analyzed. The results show that the model can reasonably describe the deformation law of an ultra-high arch dam under different water levels, conveniently and intuitively select representative measuring points and key monitoring parts, effectively reducing the analysis workload of lots of measuring points, and improve the reliability of arch dam deformation analysis.

Теория определимости в контексте информационно-коммуникационных систем

В работе обсуждается проблематика определимости и пространств отношений в исторической перспективе, обрисована роль Альфреда Тарского и Ларса Свенониуса, рассматриваются последние результаты, расширяющие полученные ранее для однородных структур, в частности на случай пополнимых вверх. Приложения включают языки описания баз данных, анализ CSP - Constraint Satisfaction Problem (обобщенной выполнимости).

An approach to finding a global optimum in constrained clustering tasks involving the assessments of several experts

An approach to solving the constrained clustering problem has been developed, based on the aggregation of data obtained as a result of evaluating the characteristics of clustered objects by several independent experts, and the analysis of alternative variants of clustering by constraint programming methods using original heuristics. Objects clusterized are represented as multisets, which makes it possible to use appropriate methods of aggregation of expert opinions. It is proposed to solve the constrained clustering problem as a constraint satisfaction problem. The main attention is paid to the issue of reducing the number and simplifying the constraints of the constraint satisfaction problem at the stage of its formalization. Within the framework of the approach, we have created: a) a method for estimating the optimal value of the objective function by hierarchical clustering of multisets, taking into account a priori constraints of the subject domain, and b) a method for generating additional constraints on the desired solution in the form of “smart tables”, based on the obtained estimate. The approach allows us to find the best partition in the problems of the class under consideration, which are characterized by a high dimension.

Nonmonotonic reasoning from conditional knowledge bases with system W

For nonmonotonic reasoning in the context of a knowledge base $\mathcal {R}$ R containing conditionals of the form If A then usually B, system P provides generally accepted axioms. Inference solely based on system P, however, is inherently skeptical because it coincides with reasoning that takes all ranking models of $\mathcal {R}$ R into account. System Z uses only the unique minimal ranking model of $\mathcal {R}$ R , and c-inference, realized via a complex constraint satisfaction problem, takes all c-representations of $\mathcal {R}$ R into account. C-representations constitute the subset of all ranking models of $\mathcal {R}$ R that are obtained by assigning non-negative integer impacts to each conditional in $\mathcal {R}$ R and summing up, for every world, the impacts of all conditionals falsified by that world. While system Z and c-inference license in general different sets of desirable entailments, the first major objective of this article is to present system W. System W fully captures and strictly extends both system Z and c-inference. Moreover, system W can be represented by a single strict partial order on the worlds over the signature of $\mathcal {R}$ R . We show that system W exhibits further inference properties worthwhile for nonmonotonic reasoning, like satisfying the axioms of system P, respecting conditional indifference, and avoiding the drowning problem. The other main goal of this article is to provide results on our investigations, underlying the development of system W, of upper and lower bounds that can be used to restrict the set of c-representations that have to be taken into account for realizing c-inference. We show that the upper bound of n − 1 is sufficient for capturing c-inference with respect to $\mathcal {R}$ R having n conditionals if there is at least one world verifying all conditionals in $\mathcal {R}$ R . In contrast to the previous conjecture that the number of conditionals in $\mathcal {R}$ R is always sufficient, we prove that there are knowledge bases requiring an upper bound of 2n− 1, implying that there is no polynomial upper bound of the impacts assigned to the conditionals in $\mathcal {R}$ R for fully capturing c-inference.

Nature-Inspired Techniques For Dynamic Constraint Satisfaction Problems

Combinatorial applications such as configuration, transportation and resource allocation, often operate under highly dynamic and unpredictable environments. In this regard, one of the main challenges is to maintain a consistent solution anytime constraints are (dynamically) added. While many solvers have been developed to tackle these applications, they often work under idealized assumptions of environmental stability. In order to address limitation, we propose a methodology, relying on nature-inspired techniques, for solving constraint problems when constraints are added dynamically. The choice for nature-inspired techniques is motivated by the fact that these are iterative algorithms, capable of maintaining a set of promising solutions, at each iteration. Our methodology takes advantage of these two properties, as follows. We first solve the initial constraint problem and save the final state (and the related population) after obtaining a consistent solution. This saved context will then be used as a resume point for finding, in an incremental manner, new solutions to subsequent variants of the problem, anytime new constraints are added. More precisely, once a solution is found, we resume from the current state to search for a new one (if the old solution is no longer feasible), when new constraints are added. This can be seen as an optimization problem where we look for a new feasible solution satisfying old and new constraints, while minimizing the differences with the solution of the previous problem, in sequence. This latter objective ensures to find the least disruptive solution, as this is very important in many applications including scheduling, planning and timetabling. Following on our proposed methodology, we have developed the dynamic variant of several nature-inspired techniques to tackle dynamic constraint problems. Constraint problems are represented using the well-known Constraint Satisfaction Problem (CSP) paradigm. Dealing with constraint additions in a dynamic environment can then be expressed as a series of static CSPs, each resulting from a change in the previous one by adding new constraints. This sequence of CSPs is called the Dynamic CSP (DCSP). To assess the performance of our proposed methodology, we conducted several experiments on randomly generated DCSP instances, following the RB model. The results of the experiments are reported and discussed.

Reactive Planning-Driven Approach to Online UAVs Mission Rerouting and Rescheduling

The presented problem concerns the route planning of a UAV fleet carrying out deliveries to spatially dispersed customers in a highly dynamic and unpredictable environment within a specified timeframe. The developed model allows for predictive (i.e., taking into account forecasted changing weather conditions) and reactive (i.e., enabling contingency UAVs rerouting) delivery mission planning (i.e., NP-hard problem) in terms of the constraint satisfaction problem. Due to the need to implement an emergency return of the UAV to the base or handling ad hoc ordered deliveries, sufficient conditions have been developed. Checking that these conditions are met allows cases to be eliminated if they do not guarantee acceptable solutions, thereby allowing the calculations to be sped up. The experiments carried out showed the usefulness of the proposed approach in DSS-based contingency planning of the UAVs’ mission performed in a dynamic environment.

Improved Algorithms for Allen's Interval Algebra: a Dynamic Programming Approach

The constraint satisfaction problem (CSP) is an important framework in artificial intelligence used to model e.g. qualitative reasoning problems such as Allen's interval algebra A. There is strong practical incitement to solve CSPs as efficiently as possible, and the classical complexity of temporal CSPs, including A, is well understood. However, the situation is more dire with respect to running time bounds of the form O(f(n)) (where n is the number of variables) where existing results gives a best theoretical upper bound 2^O(n * log n) which leaves a significant gap to the best (conditional) lower bound 2^o(n). In this paper we narrow this gap by presenting two novel algorithms for temporal CSPs based on dynamic programming. The first algorithm solves temporal CSPs limited to constraints of arity three in O(3^n) time, and we use this algorithm to solve A in O((1.5922n)^n) time. The second algorithm tackles A directly and solves it in O((1.0615n)^n), implying a remarkable improvement over existing methods since no previously published algorithm belongs to O((cn)^n) for any c. We also extend the latter algorithm to higher dimensions box algebras where we obtain the first explicit upper bound.

A Hybrid Knowledge-Based Recommender for Product-Service Systems Mass Customization

Manufacturers today compete to offer not only products, but products accompanied by services, which are referred to as product-service systems (PSSs). PSS mass customization is defined as the production of products and services to meet the needs of individual customers with near-mass-production efficiency. In the context of the PSS mass customization environment, customers are overwhelmed by a plethora of previously customized PSS variants. As a result, finding a PSS variant that is precisely aligned with the customer’s needs is a cognitive task that customers will be unable to manage effectively. In this paper, we propose a hybrid knowledge-based recommender system that assists customers in selecting previously customized PSS variants from a wide range of available ones. The recommender system (RS) utilizes ontologies for capturing customer requirements, as well as product-service and production-related knowledge. The RS follows a hybrid recommendation approach, in which the problem of selecting previously customized PSS variants is encoded as a constraint satisfaction problem (CSP), to filter out PSS variants that do not satisfy customer needs, and then uses a weighted utility function to rank the remaining PSS variants. Finally, the RS offers a list of ranked PSS variants that can be scrutinized by the customer. In this study, the proposed recommendation approach was applied to a real-life large-scale case study in the domain of laser machines. To ensure the applicability of the proposed RS, a web-based prototype system has been developed, realizing all the modules of the proposed RS.