We address the problem of imitation learning with multi-modal demonstrations. Instead attempting to learn all modes, we argue that in many tasks it is sufficient imitate any one them. show state-of-the-art methods such as GAIL and behavior cloning, due their choice loss function, often incorrectly interpolate between modes. Our key insight minimize right divergence learner expert state-action d...

BACKGROUND Tagging proteins is a standard method facilitating protein detection, purification or targeting. When tagging a certain protein of interest, it is challenging to predict which tag will give optimal results and will not interfere with protein folding, activity or yields. Ideally, multiple tags and positions are tested which however complicates molecular cloning and expression vector g...

Abstract No-cloning theorem forbids perfect cloning of an unknown quantum state. A universal machine (UQCM), capable producing two copies any input qubit with the optimal fidelity, is fundamental interest and has applications in information processing. This enabled by delicately tailored nonclassical correlations between copying qubits, which distinguish UQCM from a classical counterpart, but w...

BACKGROUND The advent of genomics-based technologies has revolutionized many fields of biological enquiry. However, chromosome walking or flanking sequence cloning is still a necessary and important procedure to determining gene structure. Such methods are used to identify T-DNA insertion sites and so are especially relevant for organisms where large T-DNA insertion libraries have been created,...

Background Despite successful cloning of many mammalian species, the cloning efficiency is extremely low compared to that of IVF raising the possibility of the existence of epigenetic barrier preventing successful cloning. MaterialsAndMethods Using comparative transcriptome analysis comparing transcriptomes of IVF and SCNT embryos and that of donor cells, we identified epigenetic barrier and fi...