Real-World Robotics: Learning to Plan for Robust Execution

In executing classical plans in the real world, small discrepancies between a planner's internal representations and the real world are unavoidable. These can conspire to cause real-world failures even though the planner is sound and, therefore, proves that a sequence of actions achieves the goal. Permissive planning, a machine learning extension to classical planning, is one response to this difficulty. This paper describes the permissive planning approach and presents GRASPER, a permissive planning robotic system that learns to robustly pick up novel objects.     

As Hangs the Flexible Line: Equilibrium of Masonry Arches

In 1675, English scientist Robert Hooke discovered “the true... ...manner of arches for building,” which he summarized with a single phrase: “As hangs the flexible line, so but inverted will stand the rigid arch.” In the centuries that followed, Hooke’s simple idea has been used to understand and design numerous important works. Recent research at MIT on the interactive analysis of structural forces provides new graphical tools for the understanding of arch behavior, which are useful for relating the forces and geometry of masonry structures. The key mathematical principle is the use of graphical analysis to determine possible equilibrium states.     

Iterated Phantom Induction: A Knowledge-Based Approach to Learning Control

We advance a knowledge-based learning method that allows prior domain knowledge to be effectively utilized by machine learning systems. The domain knowledge is incorporated not into the learning algorithm itself but instead affects only the training data. The domain knowledge is used to explain and then transform the actual training examples into a more informative set of imaginary, or phantom examples. These phantom examples are added to the training set; the experienced examples are discarded. A new control policy is induced from the phantom training set. This policy is then exercised, yielding additional training points, and the process repeats.We investigate the performance of this method in a stylized air-hockey domain which demands a difficult nonlinear control policy. Our experiments show that, surprisingly, an accurate policy can be learned even if the domain theory is only imprecise and approximate. We advance an interpretation which indicates that the information available from a plausible qualitative domain theory is sufficient for robust successful learning. This interpretation is used to make a number of predictions which are tested in subsequent experiments. The outcomes confirm the interpretation and the robustness of the approach.     

Report on workshop on high performance computing and communicationsfor grand challenge applications: computer vision, speech and naturallanguage processing, and artificial intelligence

The findings of a workshop, the goals of which were to identify applications, research problems, and designs of high performance computing and communications (HPCC) systems for supporting applications are discussed. In computer vision, the main scientific issues are machine learning, surface reconstruction, inverse optics and integration, model acquisition, and perception and action. In speech and natural language processing (SNLP), issues were identified statistical analysis in corpus-based speech and language understanding, search strategies for language analysis, auditory and vocal-tract modeling, integration of multiple levels of speech and language analyses, and connectionist systems. In AI, important issues that need immediate attention include the development of efficient machine learning and heuristic search methods that can adapt to different architectural configurations, and the design and construction of scalable and verifiable knowledge bases, active memories, and artificial neural networks     

Microbially Induced Cementation to Control Sand Response to Undrained Shear

Current methods to improve the engineering properties of sands are diverse with respect to methodology, treatment uniformity, cost, environmental impact, site accessibility requirements, etc. All of these methods have benefits and drawbacks, and there continues to be a need to explore new possibilities of soil improvement, particularly as suitable land for development becomes more scarce. This paper presents the results of a study in which natural microbial biological processes were used to engineer a cemented soil matrix within initially loose, collapsible sand. Microbially induced calcite precipitation (MICP) was achieved using the microorganism Bacillus pasteurii, an aerobic bacterium pervasive in natural soil deposits. The microbes were introduced to the sand specimens in a liquid growth medium amended with urea and a dissolved calcium source. Subsequent cementation treatments were passed through the specimen to increase the cementation level of the sand particle matrix. The results of both MICP- and gypsum-cemented specimens were assessed nondestructively by measuring the shear wave velocity with bender elements. A series of isotropically consolidated undrained compression (CIUC) triaxial tests indicate that the MICP-treated specimens exhibit a noncollapse strain softening shear behavior, with a higher initial shear stiffness and ultimate shear capacity than untreated loose specimens. This behavior is similar to that of the gypsum-cemented specimens, which represent typical cemented sand behavior. SEM microscopy verified formation of a cemented sand matrix with a concentration of precipitated calcite forming bonds at particle-particle contacts. X-ray compositional mapping confirmed that the observed cement bonds were comprised of calcite.     

Phenolic acids composition and antioxidant activity of canola extracts in cooked beef,chicken and pork

Crude polyphenol extracts (15 or 100 mg gallic acid equivalents (GAE)/kg meat) from canola meal reduced the formation of 2-thiobarbituric acid-reactive substances (TBARS) in pre-cooked beef (66–92%), pork (43–75%) and chicken (36–70%). The canola extract contained sinapic (99.7%), ferulic (0.28%) and p-hydroxybenzoic acids (0.07%).     

Isomerization of drying and semi-drying oils: The use of anthraquinone as a conjugation catalyst

Summary Anthraquinone and its derivatives have been found to be effective catalysts for the isomerization of drying and semi-drying oils to conjugated forms. A possible mechanism to account for the catalytic activity of these materials has been proposed. Possible commerical application of the isomerization of oils with anthraquinone was briefly discussed.     

Role of overconsolidation on sand–geomembrane interface response and material damage evolution

The behavior of high-density polyethylene (HDPE) reinforcement elements is critical to the overall performance of many geotechnical structures including landfills, retaining walls, embankments, and shallow foundations. The reinforcement elements, which may consist of manufactured sheets or strips, must be properly installed in order to transfer stresses from areas of concentrated loading to reinforcing zones. While attention is given during construction to avoid contact damage of reinforcement elements, the construction process exerts stresses on the reinforcement element that may substantially exceed the stresses acting throughout the structure's service life. This paper examines the influence of overconsolidation of HDPE reinforcement elements on interface performance through examination of its effect on sand–smooth geomembrane interfaces. A range of overconsolidation and operational normal stress values are selected based on conditions typically present in practice. Results show the peak friction coefficient to increase substantially with overconsolidation magnitude while the effect on the residual friction coefficient is minimal. The magnitude of the increase in peak friction coefficient is shown to be dependent on both the damage induced by the smooth geomembrane and the degree to which sand particles adjacent to the interface undergo damage in the form of breakage during shear. The damage induced by the smooth geomembrane and the sand particles are presented in terms of the level of overconsonsolidation, confinement stress during shear, and particle characteristics such as angularity and hardness.