Direct Execution In A High-Level Computer Architecture

Author(s):  
Yaohan Chu
2019 ◽  
Vol 4 (26) ◽  
pp. eaav3150 ◽  
Author(s):  
Miguel Lázaro-Gredilla ◽  
Dianhuan Lin ◽  
J. Swaroop Guntupalli ◽  
Dileep George

Humans can infer concepts from image pairs and apply those in the physical world in a completely different setting, enabling tasks like IKEA assembly from diagrams. If robots could represent and infer high-level concepts, then it would notably improve their ability to understand our intent and to transfer tasks between different environments. To that end, we introduce a computational framework that replicates aspects of human concept learning. Concepts are represented as programs on a computer architecture consisting of a visual perception system, working memory, and action controller. The instruction set of this cognitive computer has commands for parsing a visual scene, directing gaze and attention, imagining new objects, manipulating the contents of a visual working memory, and controlling arm movement. Inferring a concept corresponds to inducing a program that can transform the input to the output. Some concepts require the use of imagination and recursion. Previously learned concepts simplify the learning of subsequent, more elaborate concepts and create a hierarchy of abstractions. We demonstrate how a robot can use these abstractions to interpret novel concepts presented to it as schematic images and then apply those concepts in very different situations. By bringing cognitive science ideas on mental imagery, perceptual symbols, embodied cognition, and deictic mechanisms into the realm of machine learning, our work brings us closer to the goal of building robots that have interpretable representations and common sense.


2019 ◽  
Vol 2019 ◽  
pp. 1-18 ◽  
Author(s):  
Roberto Giorgi ◽  
Farnam Khalili ◽  
Marco Procaccini

Translating a system requirement into a low-level representation (e.g., register transfer level or RTL) is the typical goal of the design of FPGA-based systems. However, the Design Space Exploration (DSE) needed to identify the final architecture may be time consuming, even when using high-level synthesis (HLS) tools. In this article, we illustrate our hybrid methodology, which uses a frontend for HLS so that the DSE is performed more rapidly by using a higher level abstraction, but without losing accuracy, thanks to the HP-Labs COTSon simulation infrastructure in combination with our DSE tools (MYDSE tools). In particular, this proposed methodology proved useful to achieve an appropriate design of a whole system in a shorter time than trying to design everything directly in HLS. Our motivating problem was to deploy a novel execution model called data-flow threads (DF-Threads) running on yet-to-be-designed hardware. For that goal, directly using the HLS was too premature in the design cycle. Therefore, a key point of our methodology consists in defining the first prototype in our simulation framework and gradually migrating the design into the Xilinx HLS after validating the key performance metrics of our novel system in the simulator. To explain this workflow, we first use a simple driving example consisting in the modelling of a two-way associative cache. Then, we explain how we generalized this methodology and describe the types of results that we were able to analyze in the AXIOM project, which helped us reduce the development time from months/weeks to days/hours.


1992 ◽  
Vol 4 (3) ◽  
pp. 237-248
Author(s):  
Yoshimasa Goto ◽  

The Driving Pipeline is a driving control scheme for a mobile robot that drives the robot vehicle outdoors continuously and adaptively. Although the basic idea of the Driving Pipeline originates from a pipelined computer architecture, the Driving Pipeline adopts more complex execution management for adaptive vehicle motion. Like the pipelined computer architecture, the Driving Pipeline segments necessary computation for robot vehicle motion into several successive subprocesses and executes them on the pipelined processing modules that operate in parapel. Because of this pipelined architecture, the Driving Pipeline offers high computation performance, and then vehicle's high speed and continuous motion. Unlike the pipelined computer architecture, however, the Driving Pipeline adjusts execution cycles in order to adapt vehicle motion both to driving environment and computation resources in robot systems. For adaptive control, the Driving Pipeline introduces control parameters and defines required relations among them. Because of the explicit control scheme, the Driving Pipeline not only enables adaptive control but also analyzes the robot navigation. The Driving Pipeline illustrates mid level navigation between the driving control and the high level map navigation. Introducing this navigation layer offers more adaptability to the environment.


2005 ◽  
Vol 40 (1) ◽  
pp. 59-70 ◽  
Author(s):  
Karl-Erich Lindenschmidt ◽  
Jan Rauberg ◽  
Fred B. Hesser

Abstract This paper illustrates the coupling of water quality model components in High Level Architecture (HLA), a computer architecture for constructing distributed simulations. HLA facilitates interoperability among different simulations and simulation types and promotes reuse of simulation software modules. It was originally developed for military applications but the platform is finding increasing applicability for civilian purposes. The models from the Water Quality Analysis Simulation Program (WASP5) were implemented in HLA to extend its Monte Carlo uncertainty analysis capabilities. The models include DYNHYD (hydrodynamics), EUTRO (phytoplankton and nutrient dynamics) and TOXI (sediment and micropollutant transport). The uncertainty analysis investigated the impact of errors in the hydrodynamic parameters (weir discharge and roughness coefficients) and boundary conditions (upstream and tributary discharge) on the uncertainty in the water quality output variables. It was found that the contribution of the hydrodynamic parameter error to the water quality output uncertainty is comparable to that obtained from the error in the water quality parameters. The error in the boundary condition input data is also an important contributor to model uncertainty.


Sign in / Sign up

Export Citation Format

Share Document