Search

We focus on the self-supervised discovery of manipulation concepts that can be adapted and reassembled to address various robotic tasks. We propose that the decision to conceptualize a physical procedure should not depend on how we name it (semantics) but rather on the significance of the informativeness in its representation regarding the low-level physical state and state changes. We model manipulation concepts (discrete symbols) as generative and discriminative goals and derive metrics that can autonomously link them to meaningful sub-trajectories from noisy, unlabeled demonstrations. Specifically, we employ a trainable codebook containing encodings (concepts) capable of synthesizing the end-state of a sub-trajectory given the current state (generative informativeness). Moreover, the encoding corresponding to a particular sub-trajectory should differentiate the state within and outside it and confidently predict the subsequent action based on the gradient of its discriminative score (discriminative informativeness). These metrics, which do not rely on human annotation, can be seamlessly integrated into a VQ-VAE framework, enabling the partitioning of demonstrations into semantically consistent sub-trajectories, fulfilling the purpose of discovering manipulation concepts and the corresponding sub-goal (key) states. We evaluate the effectiveness of the learned concepts by training policies that utilize them as guidance, demonstrating superior performance compared to other baselines. Additionally, our discovered manipulation concepts compare favorably to human-annotated ones while saving much manual effort., Comment: 27 pages, 15 figures. Published as a conference paper at ICLR 2024

The measurement of adhesion energy between nanolayers and substrates holds significant importance for the design, fabrication, and stability assessment of micro-/nanoscale devices relying on nanolayers. In this study, we propose a nanowire-supported bridging method based on an optical microscope-based nanomanipulation technique to quantitatively measure the adhesion energy between nanolayers and substrates. Using this innovative approach, we conducted adhesion energy measurements between mica nanolayers and Si substrates, revealing a value of approximately 110 J/m2. Additionally, we discuss the applicable conditions of this new method. The proposed technique allows measurements in atmospheric conditions and is, in principle, applicable to all types of nanolayers and substrates. Consequently, it holds promise as a universal method for assessing adhesion energy between nanolayers and substrates, considering environmental factors such as atmosphere and roughness.