Search

We describe DeepMachining, a deep learning-based AI system for online prediction of machining errors of lathe machine operations. We have built and evaluated DeepMachining based on manufacturing data from factories. Specifically, we first pretrain a deep learning model for a given lathe machine's operations to learn the salient features of machining states. Then, we fine-tune the pretrained model to adapt to specific machining tasks. We demonstrate that DeepMachining achieves high prediction accuracy for multiple tasks that involve different workpieces and cutting tools. To the best of our knowledge, this work is one of the first factory experiments using pre-trained deep-learning models to predict machining errors of lathe machines.

Standard recognition approaches are unable to deal with novel categories at test time. Their overconfidence on the known classes makes the predictions unreliable for safety-critical applications such as healthcare or autonomous driving. Out-Of-Distribution (OOD) detection methods provide a solution by identifying semantic novelty. Most of these methods leverage a learning stage on the known data, which means training (or fine-tuning) a model to capture the concept of normality. This process is clearly sensitive to the amount of available samples and might be computationally expensive for on-board systems. A viable alternative is that of evaluating similarities in the embedding space produced by large pre-trained models without any further learning effort. We focus exactly on such a fine-tuning-free OOD detection setting. This works presents an in-depth analysis of the recently introduced relational reasoning pre-training and investigates the properties of the learned embedding, highlighting the existence of a correlation between the inter-class feature distance and the OOD detection accuracy. As the class separation depends on the chosen pre-training objective, we propose an alternative loss function to control the inter-class margin, and we show its advantage with thorough experiments., Comment: Accepted for publication at ICIAP 2023

We investigate the spin and pseudospin symmetry in the single-particle resonant states by solving the Dirac equation containing a Woods-Saxon potential with Green's function method. Taking double-magic nucleus $^{208}$Pb as an example, three spin doublets $3d$, $2h$, and $1j$ and three pseudospin doublets $3\tilde{p}$, $1\tilde{i}$, and $1\tilde{j}$ are obtained for the single-neutron resonant states. By analyzing the energy splittings, we find that the threshold effect plays an important role in resonant pseudospin doubles. Besides, there is a reversed level structure of pseudospin doublets in the continuum. Differently, all the width splittings of either the spin doublets or the pseudospin doublets are systematically positive and the splittings are very small except $1\tilde{j}$ doublet. Further studies show that the splittings of the energies and widths for the resonant (pseudo)spin doublets are independent. Besides, the similarity properties of the wave functions of the spin and pseudospin doublets still maintain well in resonant states., Comment: 10 pages, 8 figures, 1 table. arXiv admin note: text overlap with arXiv:1710.02923

Topological semimetal may have substantial applications in electronics, spintronics and quantum computation. Recently, ZrTe is predicted as a new type of topological semimetal due to coexistence of Weyl fermion and massless triply degenerate nodal points. In this work, the elastic and transport properties of ZrTe are investigated by combining the first-principles calculations and semiclassical Boltzmann transport theory. Calculated elastic constants prove mechanical stability of ZrTe, and the bulk modulus, shear modulus, Young's modulus and Possion's ratio also are calculated. It is found that spin-orbit coupling (SOC) has slightly enhanced effects on Seebeck coefficient, which along a(b) and c directions for pristine ZrTe at 300 K is 46.26 $\mu$V/K and 80.20 $\mu$V/K, respectively. By comparing the experimental electrical conductivity of ZrTe (300 K) with calculated value, the scattering time is determined for 1.59 $\times$ $10^{-14}$ s. The predicted room-temperature electronic thermal conductivity along a(b) and c directions is 2.37 $\mathrm{W m^{-1} K^{-1}}$ and 2.90 $\mathrm{W m^{-1} K^{-1}}$, respectively. The room-temperature lattice thermal conductivity is predicted as 17.56 $\mathrm{W m^{-1} K^{-1}}$ and 43.08 $\mathrm{W m^{-1} K^{-1}}$ along a(b) and c directions, showing very strong anisotropy. Calculated results show that isotope scattering produces observable effect on lattice thermal conductivity. It is noted that average room-temperature lattice thermal conductivity of ZrTe is slightly higher than that of isostructural MoP, which is due to larger phonon lifetimes and smaller Gr$\mathrm{\ddot{u}}$neisen parameters. Finally, the total thermal conductivity as a function of temperature is predicted for pristine ZrTe., Comment: 9 pages, 10 figures

It is still under debate whether the classical Einstein relation in disordered organic semiconductors is valid. We investigated Einstein relation in disordered organic semiconductors theoretically. The results show that, the classic Einstein relation deviate dramatically with disorder and electric field, even in the case of thermal equilibrium.

Within the relativistic Hartree-Fock-Bogoliubov (RHFB) theory, the structure properties of Carbon isotopes are systematically studied. In order to reproduce the experiment data, we take the finite-range Gogny D1S with a strength factor $f$ as the pairing force. The self-consistent RHFB calculations with density-dependent meson-nucleon couplings indicate the single-neutron halo structures in both $^{17}$C and $^{19}$C, whereas the two-neutron halo in $^{22}$C is not well supported. It is also found that close to the neutron drip line there exists distinct odd-even staggering on neutron radii, which is tightly related with the blocking effects and correspondingly the blocking effect plays a significant role in halo formation., Comment: 8 pages, 5 figures, 5 table

This article concerns the application of bootstrap methodology to construct a likelihood-based confidence region for operating conditions associated with the maximum of a response surface constrained to a specified region. Unlike classical methods based on the stationary point, proper interpretation of this confidence region does not depend on unknown model parameters. In addition, the methodology does not require the assumption of normally distributed errors. The approach is demonstrated for concave-down and saddle system cases in two dimensions. Simulation studies were performed to assess the coverage probability of these regions., Comment: Submitted to the Electronic Journal of Statistics (http://www.i-journals.org/ejs/) by the Institute of Mathematical Statistics (http://www.imstat.org)