In this paper, we propose a no-reference (NR) image quality assessment (IQA) method via feature level pseudo-reference (PR) hallucination. The proposed quality assessment framework is grounded on the prior models of natural image statistical behavior
s and rooted in the view that the perceptually meaningful features could be well exploited to characterize the visual quality. Herein, the PR features from the distorted images are learned by a mutual learning scheme with the pristine reference as the supervision, and the discriminative characteristics of PR features are further ensured with the triplet constraints. Given a distorted image for quality inference, the feature level disentanglement is performed with an invertible neural layer for final quality prediction, leading to the PR and the corresponding distortion features for comparison. The effectiveness of our proposed method is demonstrated on four popular IQA databases, and superior performance on cross-database evaluation also reveals the high generalization capability of our method. The implementation of our method is publicly available on https://github.com/Baoliang93/FPR.
In this paper, we consider the development of efficient numerical methods for linear transport equations with random parameters and under the diffusive scaling. We extend to the present case the bi-fidelity stochastic collocation method introduced in
[33,50,51]. For the high-fidelity transport model, the asymptotic-preserving scheme [29] is used for each stochastic sample. We employ the simple two-velocity Goldstein-Taylor equation as low-fidelity model to accelerate the convergence of the uncertainty quantification process. The choice is motivated by the fact that both models, high fidelity and low fidelity, share the same diffusion limit. Speed-up is achieved by proper selection of the collocation points and relative approximation of the high-fidelity solution. Extensive numerical experiments are conducted to show the efficiency and accuracy of the proposed method, even in non diffusive regimes, with empirical error bound estimations as studied in [16].
We study the problem of safe offline reinforcement learning (RL), the goal is to learn a policy that maximizes long-term reward while satisfying safety constraints given only offline data, without further interaction with the environment. This proble
m is more appealing for real world RL applications, in which data collection is costly or dangerous. Enforcing constraint satisfaction is non-trivial, especially in offline settings, as there is a potential large discrepancy between the policy distribution and the data distribution, causing errors in estimating the value of safety constraints. We show that naive approaches that combine techniques from safe RL and offline RL can only learn sub-optimal solutions. We thus develop a simple yet effective algorithm, Constraints Penalized Q-Learning (CPQ), to solve the problem. Our method admits the use of data generated by mixed behavior policies. We present a theoretical analysis and demonstrate empirically that our approach can learn robustly across a variety of benchmark control tasks, outperforming several baselines.
The low degradability of common polymers composed of light elements, results in a serious impact on the environment, which has become an urgent problem to be solved. As the reverse process of monomer polymerization, what deviates degradation from the
idealized sequential depolymerization process, thereby bringing strange degradation products or even hindering further degradation? This is a key issue at the atomic level that must be addressed. Herein, we reveal that hydrogen atom transfer (HAT) during degradation, which is usually attributed to the thermal effect, unexpectedly exhibits a strong high-temperature tunnelling effect. This gives a possible answer to the above question. High-precision first-principles calculations show that, in various possible HAT pathways, lower energy barrier and stronger tunnelling effect make the HAT reaction related to the active end of the polymer occur more easily. In particular, although the energy barrier of the HAT reaction is only of 0.01 magnitude different from depolymerization, the tunnelling probability of the former can be 14~32 orders of magnitude greater than that of the latter. Furthermore, chain scission following HAT will lead to a variety of products other than monomers. Our work highlights that quantum tunnelling may be an important source of uncertainty in degradation and will provide a direction for regulating the polymer degradation process.
Many studies have revealed that confined water chain flipping is closely related to the spatial size and even quantum effects of the confinement environment. Here, we show that these are not the only factors that affect the flipping process of a conf
ined water chain. First-principles calculations and analyses confirm that quantum tunnelling effects from the water chain itself, especially resonant tunnelling, enhance the hydrogen bond rotation process. Importantly, resonant tunnelling can result in tunnelling rotation of hydrogen bonds with a probability close to 1 with only 0.597 eV provided energy. Compared to sequential tunnelling, resonant tunnelling dominants water chain flipping at temperatures up to 20 K higher. Additionally, the ratio of the resonant tunnelling probability to the thermal disturbance probability at 200 K is at least ten times larger than that of sequential tunnelling, which further illustrates the enhancement of hydrogen bond rotation brought about by resonant tunnelling.
Federated learning is an emerging distributed machine learning framework for privacy preservation. However, models trained in federated learning usually have worse performance than those trained in the standard centralized learning mode, especially w
hen the training data are not independent and identically distributed (Non-IID) on the local devices. In this survey, we pro-vide a detailed analysis of the influence of Non-IID data on both parametric and non-parametric machine learning models in both horizontal and vertical federated learning. In addition, cur-rent research work on handling challenges of Non-IID data in federated learning are reviewed, and both advantages and disadvantages of these approaches are discussed. Finally, we suggest several future research directions before concluding the paper.
Purpose: Recent studies suggest ultra-high dose rate (FLASH) irradiation can spare normal tissues from radiotoxicity, while efficiently controlling the tumor, and this is known as the FLASH effect. This study performed theoretical analyses about the
impact of radiolytic oxygen depletion (ROD) on the cellular responses after FLASH irradiation. Methods: Monte Carlo simulation was used to model the ROD process, determine the DNA damage, and calculate the amount of oxygen depleted (LROD) during FLASH exposure. A mathematical model was applied to analyze oxygen tension (pO2) distribution in human tissues and the recovery of pO2 after FLASH irradiation. DNA damage and cell survival fractions (SFs) after FLASH irradiation were calculated. The impact of initial cellular pO2, FLASH pulse number, pulse interval, and radiation quality of the source particles on ROD and subsequent cellular responses were systematically evaluated. Results: The simulated electron LROD range was 0.38-0.43 {mu}M/Gy when pO2 ranged from 7.5-160 mmHg. The calculated DNA damage and SFs show that radioprotective effect is only evident in cells with a lower pO2. Different irradiation setups alter the cellular responses by modifying the pO2. Single pulse delivery or multi-pulse delivery with pulse intervals shorter than 10-50 ms resulted in fewer DNA damages and higher SFs. Source particles with a low radiation quality have a higher capacity to deplete oxygen, and thus, lead to a more conspicuous radioprotective effect. Conclusions: The FLASH radioprotective effect due to ROD may only be observed in cells with a low pO2. Single pulse delivery or multi-pulse delivery with short pulse intervals are suggested for FLASH irradiation to avoid oxygen tension recovery during pulse intervals. Source particles with low radiation quality are preferred for their conspicuous radioprotective effects.
We explore the multi-faceted important features of turbulence (e.g., anisotropy, dispersion, diffusion) in the three-dimensional (3D) wavenumber domain ($k_parallel$, $k_{perp,1}$, $k_{perp,2}$), by employing the k-filtering technique to the high-qua
lity measurements of fields and particles from the MMS multi-spacecraft constellation. We compute the 3D power spectral densities (PSDs) of magnetic and electric fluctuations (marked as $rm{PSD}(delta mathbf{B}(mathbf{k}))$ and $rm{PSD}(delta mathbf{E}_{langlemathbf{v}_mathrm{i}rangle}(mathbf{k}))$), both of which show a prominent spectral anisotropy in the sub-ion range. We give the first 3D image of the bifurcation between power spectra of the electric and magnetic fluctuations, by calculating the ratio between $rm{PSD}(delta mathbf{E}_{ langlemathbf{v}_mathrm{i}rangle}(mathbf{k}))$ and $rm{PSD}(delta mathbf{B}(mathbf{k}))$, the distribution of which is related to the non-linear dispersion relation. We also compute the ratio between electric spectra in different reference frames defined by the ion bulk velocity, that is $mathrm{PSD}(delta{mathbf{E}_{mathrm{local} mathbf{v}_mathrm{i}}})/mathrm{PSD}(delta{mathbf{E}_{ langlemathbf{v}_mathrm{i}rangle}})$, to visualize the turbulence ion diffusion region (T-IDR) in wavenumber space. The T-IDR has an anisotropy and a preferential direction of wavevectors, which is generally consistent with the plasma wave theory prediction based on the dominance of kinetic Alfven waves (KAW). This work manifests the worth of the k-filtering technique in diagnosing turbulence comprehensively, especially when the electric field is involved.
Offline reinforcement learning (RL) enables learning policies using pre-collected datasets without environment interaction, which provides a promising direction to make RL useable in real-world systems. Although recent offline RL studies have achieve
d much progress, existing methods still face many practical challenges in real-world system control tasks, such as computational restriction during agent training and the requirement of extra control flexibility. Model-based planning framework provides an attractive solution for such tasks. However, most model-based planning algorithms are not designed for offline settings. Simply combining the ingredients of offline RL with existing methods either provides over-restrictive planning or leads to inferior performance. We propose a new light-weighted model-based offline planning framework, namely MOPP, which tackles the dilemma between the restrictions of offline learning and high-performance planning. MOPP encourages more aggressive trajectory rollout guided by the behavior policy learned from data, and prunes out problematic trajectories to avoid potential out-of-distribution samples. Experimental results show that MOPP provides competitive performance compared with existing model-based offline planning and RL approaches, and allows easy adaptation to varying objectives and extra constraints.
It is a long-standing question to discover causal relations among a set of variables in many empirical sciences. Recently, Reinforcement Learning (RL) has achieved promising results in causal discovery from observational data. However, searching the
space of directed graphs and enforcing acyclicity by implicit penalties tend to be inefficient and restrict the existing RL-based method to small scale problems. In this work, we propose a novel RL-based approach for causal discovery, by incorporating RL into the ordering-based paradigm. Specifically, we formulate the ordering search problem as a multi-step Markov decision process, implement the ordering generating process with an encoder-decoder architecture, and finally use RL to optimize the proposed model based on the reward mechanisms designed for~each ordering. A generated ordering would then be processed using variable selection to obtain the final causal graph. We analyze the consistency and computational complexity of the proposed method, and empirically show that a pretrained model can be exploited to accelerate training. Experimental results on both synthetic and real data sets shows that the proposed method achieves a much improved performance over existing RL-based method.
