skip to main content

Search for: All records

Creators/Authors contains: "Sharma, M."

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. To perform manipulation tasks in the real world, robots need to operate on objects with various shapes, sizes and without access to geometric models. To achieve this it is often infeasible to train monolithic neural network policies across such large variations in object properties. Towards this generalization challenge, we propose to learn modular task policies which compose object-centric task-axes controllers. These task-axes controllers are parameterized by properties associated with underlying objects in the scene. We infer these controller parameters directly from visual input using multi- view dense correspondence learning. Our overall approach provides a simple and yet powerful framework formore »learning manipulation tasks. We empirically evaluate our approach on 3 different manipulation tasks and show its ability to generalize to large variance in object size, shape and geometry.« less
  2. Manipulation tasks can often be decomposed into multiple subtasks performed in parallel, e.g., sliding an object to a goal pose while maintaining con- tact with a table. Individual subtasks can be achieved by task-axis controllers defined relative to the objects being manipulated, and a set of object-centric controllers can be combined in an hierarchy. In prior works, such combinations are defined manually or learned from demonstrations. By contrast, we propose using reinforcement learning to dynamically compose hierarchical object-centric controllers for manipulation tasks. Experiments in both simulation and real world show how the proposed approach leads to improved sample efficiency, zero-shotmore »generalization to novel test environments, and simulation-to-reality transfer with- out fine-tuning.« less
  3. Free, publicly-accessible full text available August 1, 2022
  4. Free, publicly-accessible full text available September 1, 2022
  5. Free, publicly-accessible full text available September 1, 2022
  6. Free, publicly-accessible full text available August 1, 2022
  7. Abstract The coherent photoproduction of $$\mathrm{J}/\psi $$ J / ψ and $${\uppsi '}$$ ψ ′ mesons was measured in ultra-peripheral Pb–Pb collisions at a center-of-mass energy $$\sqrt{s_{\mathrm {NN}}}~=~5.02$$ s NN = 5.02  TeV  with the ALICE detector. Charmonia are detected in the central rapidity region for events where the hadronic interactions are strongly suppressed. The $$\mathrm{J}/\psi $$ J / ψ is reconstructed using the dilepton ( $$l^{+} l^{-}$$ l + l - ) and proton–antiproton decay channels, while for the $${\uppsi '}$$ ψ ′   the dilepton and the $$l^{+} l^{-} \pi ^{+} \pi ^{-}$$ l + l - πmore »+ π - decay channels are studied. The analysis is based on an event sample corresponding to an integrated luminosity of about 233 $$\mu {\mathrm{b}}^{-1}$$ μ b - 1 . The results are compared with theoretical models for coherent $$\mathrm{J}/\psi $$ J / ψ and $${\uppsi '}$$ ψ ′ photoproduction. The coherent cross section is found to be in a good agreement with models incorporating moderate nuclear gluon shadowing of about 0.64 at a Bjorken- x of around $$6\times 10^{-4}$$ 6 × 10 - 4 , such as the EPS09 parametrization, however none of the models is able to fully describe the rapidity dependence of the coherent $$\mathrm{J}/\psi $$ J / ψ cross section including ALICE measurements at forward rapidity. The ratio of $${\uppsi '}$$ ψ ′ to $$\mathrm{J}/\psi $$ J / ψ coherent photoproduction cross sections was also measured and found to be consistent with the one for photoproduction off protons.« less
    Free, publicly-accessible full text available August 1, 2022