arXiv subjects

Jie Song

Publications and source records attributed to Jie Song.

At least 235 records · Page 13Linked to original sources

arXiv2020

Convolutional Autoencoders for Human Motion Infilling

In this paper we propose a convolutional autoencoder to address the problem of motion infilling for 3D human motion data. Given a start and end sequence, motion infilling aims to complete the missing gap in between, such that the filled in poses plausibly forecast the start sequence and naturally transition into the end sequence. To this end, we propose a single, end-to-end trainable convolutional autoencoder. We show that a single model can be used to create natural transitions between different types of activities. Furthermore, our method is not only able to fill in entire missing frames, but it can also be used to complete gaps where partial poses are available (e.g. from end effectors), or to clean up other forms of noise (e.g. Gaussian). Also, the model can fill in an arbitrary number of gaps that potentially vary in length. In addition, no further post-processing on the model's outputs is necessary such as smoothing or closing discontinuities at the end of the gap. At the heart of our approach lies the idea to cast motion infilling as an inpainting problem and to train a convolutional de-noising autoencoder on image-like representations of motion sequences. At training time, blocks of columns are removed from such images and we ask the model to fill in the gaps. We demonstrate the versatility of the approach via a number of complex motion sequences and report on thorough evaluations performed to better understand the capabilities and limitations of the proposed approach.

cs.CV↗

arXiv2020

Robust generation of logical qubit singlet states with reverse engineering and optimal control with spin qubits

A protocol is proposed to generate singlet states of three logical qubits constructed by pairs of spins. Single and multiple operations of logical qubits are studied for the construction of an effective Hamiltonian, with which robust control fields are derived with invariant-based reverse engineering and optimal control. Moreover, systematic errors are further compensated by periodic modulation for better robustness. Furthermore, resistance to decoherence of the protocol is also shown with numerical simulations. Therefore, the protocol may provide useful perspectives for generations of logical qubit entanglement in spin systems.

quant-ph↗

arXiv2020

Human Body Model Fitting by Learned Gradient Descent

We propose a novel algorithm for the fitting of 3D human shape to images. Combining the accuracy and refinement capabilities of iterative gradient-based optimization techniques with the robustness of deep neural networks, we propose a gradient descent algorithm that leverages a neural network to predict the parameter update rule for each iteration. This per-parameter and state-aware update guides the optimizer towards a good solution in very few steps, converging in typically few steps. During training our approach only requires MoCap data of human poses, parametrized via SMPL. From this data the network learns a subspace of valid poses and shapes in which optimization is performed much more efficiently. The approach does not require any hard to acquire image-to-3D correspondences. At test time we only optimize the 2D joint re-projection error without the need for any further priors or regularization terms. We show empirically that this algorithm is fast (avg. 120ms convergence), robust to initialization and dataset, and achieves state-of-the-art results on public evaluation datasets including the challenging 3DPW in-the-wild benchmark (improvement over SMPLify 45%) and also approaches using image-to-3D correspondences

cs.CV↗

arXiv2020

Category Level Object Pose Estimation via Neural Analysis-by-Synthesis

Many object pose estimation algorithms rely on the analysis-by-synthesis framework which requires explicit representations of individual object instances. In this paper we combine a gradient-based fitting procedure with a parametric neural image synthesis module that is capable of implicitly representing the appearance, shape and pose of entire object categories, thus rendering the need for explicit CAD models per object instance unnecessary. The image synthesis network is designed to efficiently span the pose configuration space so that model capacity can be used to capture the shape and local appearance (i.e., texture) variations jointly. At inference time the synthesized images are compared to the target via an appearance based loss and the error signal is backpropagated through the network to the input parameters. Keeping the network parameters fixed, this allows for iterative optimization of the object pose, shape and appearance in a joint manner and we experimentally show that the method can recover orientation of objects with high accuracy from 2D images alone. When provided with depth measurements, to overcome scale ambiguities, the method can accurately recover the full 6DOF pose successfully.

cs.CV↗

arXiv2020

Sparse Coding Driven Deep Decision Tree Ensembles for Nuclear Segmentation in Digital Pathology Images

In this paper, we propose an easily trained yet powerful representation learning approach with performance highly competitive to deep neural networks in a digital pathology image segmentation task. The method, called sparse coding driven deep decision tree ensembles that we abbreviate as ScD2TE, provides a new perspective on representation learning. We explore the possibility of stacking several layers based on non-differentiable pairwise modules and generate a densely concatenated architecture holding the characteristics of feature map reuse and end-to-end dense learning. Under this architecture, fast convolutional sparse coding is used to extract multi-level features from the output of each layer. In this way, rich image appearance models together with more contextual information are integrated by learning a series of decision tree ensembles. The appearance and the high-level context features of all the previous layers are seamlessly combined by concatenating them to feed-forward as input, which in turn makes the outputs of subsequent layers more accurate and the whole model efficient to train. Compared with deep neural networks, our proposed ScD2TE does not require back-propagation computation and depends on less hyper-parameters. ScD2TE is able to achieve a fast end-to-end pixel-wise training in a layer-wise manner. We demonstrated the superiority of our segmentation technique by evaluating it on the multi-disease state and multi-organ dataset where consistently higher performances were obtained for comparison against several state-of-the-art deep learning methods such as convolutional neural networks (CNN), fully convolutional networks (FCN), etc.

cs.CV↗

arXiv2020

Noise-resistant phase gates with amplitude modulation

We propose a simple scheme for implementing fast arbitrary phase gates and employ pulse modulation to improve the gate robustness against different sources of noise. Parametric driving of a cavity is introduced to induce Rabi interactions between the cavity and qutrits, and then a two-qubit arbitrary phase gate is constructed by designing proper logical states. On this basis, we implement amplitude-shaped gates to obtain enhanced resilience to control errors in gate time and frequency detuning. By virtue of specifically designed logical states, the scheme displays intrinsic resistance to the energy relaxations of qutrits. Furthermore, we show that the gate robustness against cavity decay can be enhanced significantly with amplitude modulation.

quant-ph↗

arXiv2020

Heralded atomic nonadiabatic holonomic quantum computation with Rydberg blockade

We propose a protocol to realize atomic nonadiabatic holonomic quantum computation (NHQC) with two computational atoms and an auxiliary atom. Dynamics of the system is analyzed in the regime of Rydberg blockade, and robust laser pulses are designed via reverse engineering, so that quantum gates can be easily realized with high fidelities. In addition, we also study the evolution suffering from dissipation with a master equation. The result indicates that decays of atoms can be heralded by measuring the state of the auxiliary atom, and nearly perfect unitary evolution can be obtained if the auxiliary atom remains in its Rydberg state. Therefore, the protocol may be helpful to realize NHQC in dissipative environment.

quant-ph↗

arXiv2020

Effective discrimination of chiral molecules in a cavity

We present a scheme to realize precise discrimination of chiral molecules in a cavity. Assisted by additional laser pulses, cavity fields can evolve to different coherence states with contrary-sign displacements according to the handedness of molecules. Consequently, the handedness of molecules can be read out with homodyne measurement on the cavity, and the successful probability is nearly unity without very strong cavity fields. Numerical results show that the scheme is insensitive to errors, noise and decoherence. Therefore, the scheme may provide helpful perspectives for accurate discrimination of chiral molecules.

quant-ph↗

arXiv2020

Impression Space from Deep Template Network

It is an innate ability for humans to imagine something only according to their impression, without having to memorize all the details of what they have seen. In this work, we would like to demonstrate that a trained convolutional neural network also has the capability to "remember" its input images. To achieve this, we propose a simple but powerful framework to establish an {\emph{Impression Space}} upon an off-the-shelf pretrained network. This network is referred to as the {\emph{Template Network}} because its filters will be used as templates to reconstruct images from the impression. In our framework, the impression space and image space are bridged by a layer-wise encoding and iterative decoding process. It turns out that the impression space indeed captures the salient features from images, and it can be directly applied to tasks such as unpaired image translation and image synthesis through impression matching without further network training. Furthermore, the impression naturally constructs a high-level common space for different data. Based on this, we propose a mechanism to model the data relations inside the impression space, which is able to reveal the feature similarity between images. Our code will be released.

cs.CV↗

arXiv2020

Turbulence on the Global Economy influenced by Artificial Intelligence and Foreign Policy Inefficiencies

It is said that Data and Information are the new oil. One, who handles the data, handles the emerging future of the global economy. Complex algorithms and intelligence-based filter programs are utilized to manage, store, handle and maneuver vast amounts of data for the fulfillment of specific purposes. This paper seeks to find the bridge between artificial intelligence and its impact on the international policy implementation in the light of geopolitical influence, global economy and the future of labor markets. We hypothesize that the distortion in the labor markets caused by artificial intelligence can be mitigated by a collaborative international foreign policy on the deployment of AI in the industrial circles. We, in this paper, then proceed to propose a disposition for the essentials of AI-based foreign policy and implementation, while asking questions such as 'could AI become the real Invisible Hand discussed by economists?'.

cs.CY↗

arXiv2020

Two-path interference for enantiomer-selective state transfer of chiral molecules

With a microwave-regime cyclic three-state configuration, an enantiomer-selective state transfer~(ESST) is carried out through the two-path interference between a direct one-photon coupling and an effective two-photon coupling. The $π$-phase difference in the one-photon process between two enantiomers makes the interference constructive for one enantiomer but destructive for the other. Therefore only one enantiomer is excited into a higher rotational state while the other remains in the ground state. The scheme is of flexibility in the pulse waveforms and the time order of two paths. We simulate the scheme in a sample of cyclohexylmethanol~(C$_7$H$_{14}$O) molecules. Simulative results show the robust and high-fidelity ESST can be obtained when experimental concerns are considered. Finally, we propose to employ the finished ESST in implementing enantio-separation and determining enantiomeric excess.

quant-ph↗

arXiv2020

DEPARA: Deep Attribution Graph for Deep Knowledge Transferability

Exploring the intrinsic interconnections between the knowledge encoded in PRe-trained Deep Neural Networks (PR-DNNs) of heterogeneous tasks sheds light on their mutual transferability, and consequently enables knowledge transfer from one task to another so as to reduce the training effort of the latter. In this paper, we propose the DEeP Attribution gRAph (DEPARA) to investigate the transferability of knowledge learned from PR-DNNs. In DEPARA, nodes correspond to the inputs and are represented by their vectorized attribution maps with regards to the outputs of the PR-DNN. Edges denote the relatedness between inputs and are measured by the similarity of their features extracted from the PR-DNN. The knowledge transferability of two PR-DNNs is measured by the similarity of their corresponding DEPARAs. We apply DEPARA to two important yet under-studied problems in transfer learning: pre-trained model selection and layer selection. Extensive experiments are conducted to demonstrate the effectiveness and superiority of the proposed method in solving both these problems. Code, data and models reproducing the results in this paper are available at \url{https://github.com/zju-vipa/DEPARA}.

cs.CV↗

arXiv2020

Effective Rabi dynamics of Rydberg atoms and robust high-fidelity quantum gates with a resonant amplitude-modulation field

With a resonant amplitude-modulation field on two Rydberg atoms, we propose a Rydberg antiblockade (RAB) regime, where the Rabi oscillation between collective ground and excited states is induced. A controlled-Z gate can be yielded through a Rabi cycle. Further, several common issues of the RAB gates are solved by modifying the parameter relation. The gate fidelity and the gate robustness against the control error are enhanced with a shaped pulse. The requirement of control precision of the Rydberg-Rydberg interaction strength is relaxed. In addition, the atomic excitation is restrained and therefore the gate robustness against the atomic decay is enhanced.

quant-ph↗

arXiv2020

Data-Free Adversarial Distillation

Knowledge Distillation (KD) has made remarkable progress in the last few years and become a popular paradigm for model compression and knowledge transfer. However, almost all existing KD algorithms are data-driven, i.e., relying on a large amount of original training data or alternative data, which is usually unavailable in real-world scenarios. In this paper, we devote ourselves to this challenging problem and propose a novel adversarial distillation mechanism to craft a compact student model without any real-world data. We introduce a model discrepancy to quantificationally measure the difference between student and teacher models and construct an optimizable upper bound. In our work, the student and the teacher jointly act the role of the discriminator to reduce this discrepancy, when a generator adversarially produces some "hard samples" to enlarge it. Extensive experiments demonstrate that the proposed data-free method yields comparable performance to existing data-driven methods. More strikingly, our approach can be directly extended to semantic segmentation, which is more complicated than classification, and our approach achieves state-of-the-art results. Code and pretrained models are available at https://github.com/VainF/Data-Free-Adversarial-Distillation.

cs.LG↗

arXiv2019

Deep Model Transferability from Attribution Maps

Exploring the transferability between heterogeneous tasks sheds light on their intrinsic interconnections, and consequently enables knowledge transfer from one task to another so as to reduce the training effort of the latter. In this paper, we propose an embarrassingly simple yet very efficacious approach to estimating the transferability of deep networks, especially those handling vision tasks. Unlike the seminal work of taskonomy that relies on a large number of annotations as supervision and is thus computationally cumbersome, the proposed approach requires no human annotations and imposes no constraints on the architectures of the networks. This is achieved, specifically, via projecting deep networks into a model space, wherein each network is treated as a point and the distances between two points are measured by deviations of their produced attribution maps. The proposed approach is several-magnitude times faster than taskonomy, and meanwhile preserves a task-wise topological structure highly similar to the one obtained by taskonomy. Code is available at https://github.com/zju-vipa/TransferbilityFromAttributionMaps.

cs.CV↗

arXiv2019

Robust and highly-efficient discrimination of chiral molecules through three-mode parallel paths

We propose to discriminate chiral molecules by combining one- and two-photon processes in a closed-loop configuration. The one-photon-coupling intrinsic π-phase difference between two enantiomers leads to their different superposition states, which is then followed by a two-photon process through three-mode parallel paths (3MPPs), enabling the discrimination of enantiomers by inducing their entirely-different population distributions. The 3MPPs are constructed by "chosen paths", a method of shortcuts to adiabaticity (STA), exhibiting a fast two-photon process. As an example, we propose to perform the scheme in 1, 2-propanediol molecules, which shows relatively robust and highly-efficient results under considering the experimental issues concerning unwanted transitions, imperfect initial state, pulse shaping, control errors and the effect of energy relaxations. The present work may provide help for laboratory researchers in a robust separation of chiral molecules.

quant-ph↗

arXiv2019

Monocular Neural Image Based Rendering with Continuous View Control

We present an approach that learns to synthesize high-quality, novel views of 3D objects or scenes, while providing fine-grained and precise control over the 6-DOF viewpoint. The approach is self-supervised and only requires 2D images and associated view transforms for training. Our main contribution is a network architecture that leverages a transforming auto-encoder in combination with a depth-guided warping procedure to predict geometrically accurate unseen views. Leveraging geometric constraints renders direct supervision via depth or flow maps unnecessary. If large parts of the object are occluded in the source view, a purely learning based prior is used to predict the values for dis-occluded pixels. Our network furthermore predicts a per-pixel mask, used to fuse depth-guided and pixel-based predictions. The resulting images reflect the desired 6-DOF transformation and details are preserved. We thoroughly evaluate our architecture on synthetic and real scenes and under fine-grained and fixed-view settings. Finally, we demonstrate that the approach generalizes to entirely unseen images such as product images downloaded from the internet.

cs.CV↗

arXiv2019

Customizing Student Networks From Heterogeneous Teachers via Adaptive Knowledge Amalgamation

A massive number of well-trained deep networks have been released by developers online. These networks may focus on different tasks and in many cases are optimized for different datasets. In this paper, we study how to exploit such heterogeneous pre-trained networks, known as teachers, so as to train a customized student network that tackles a set of selective tasks defined by the user. We assume no human annotations are available, and each teacher may be either single- or multi-task. To this end, we introduce a dual-step strategy that first extracts the task-specific knowledge from the heterogeneous teachers sharing the same sub-task, and then amalgamates the extracted knowledge to build the student network. To facilitate the training, we employ a selective learning scheme where, for each unlabelled sample, the student learns adaptively from only the teacher with the least prediction ambiguity. We evaluate the proposed approach on several datasets and experimental results demonstrate that the student, learned by such adaptive knowledge amalgamation, achieves performances even better than those of the teachers.

cs.CV↗