Optically Probing Unconventional Superconductivity in Atomically Thin Bi2Sr2Ca0.92Y0.08Cu2O8+δ.

Atomically thin cuprates exhibiting a superconducting phase transition temperature similar to that of the bulk have recently been realized, although the device fabrication remains a challenge and limits the potential for many novel studies and applications. Here, we use an optical pump-probe approach to noninvasively study the unconventional superconductivity in atomically thin Bi2Sr2Ca0.92Y0.08Cu2O8+δ (Y-Bi2212). Apart from finding an optical response due to the superconducting phase transition that is similar to that of bulk Y-Bi2212, we observe that the sign and amplitude of the pump-probe signal in atomically thin flakes vary significantly in different dielectric environments depending on the nature of the optical excitation. By exploiting the spatial resolution of the optical probe, we uncover the exceptional sensitivity of monolayer Y-Bi2212 to the environment. Our results provide the first optical evidence for the intralayer nature of the superconducting condensate in Bi2212 and highlight the role of double-sided encapsulation in preserving superconductivity in atomically thin cuprates.

Polaritonic Chern Insulators in Monolayer Semiconductors.

Systems with strong light-matter interaction open up new avenues for studying topological phases of matter. Examples include exciton polaritons, mixed light-matter quasiparticles, where the topology of the polaritonic band structure arises from the collective coupling between matter wave and optical fields strongly confined in periodic dielectric structures. Distinct from light-matter interaction in a uniform environment, the spatially varying nature of the optical fields leads to a fundamental modification of the well-known optical selection rules, which were derived under the plane wave approximation. Here we identify polaritonic Chern insulators by coupling valley excitons in transition metal dichalcogenides to photonic Bloch modes in a dielectric photonic crystal slab. We show that polaritonic Dirac points, which are markers for topological phase transition points, can be constructed from the collective coupling between valley excitons and photonic Dirac cones in the presence of both time-reversal and inversion symmetries. Lifting exciton valley degeneracy by breaking time-reversal symmetry leads to gapped polaritonic bands with nonzero Chern numbers. Through numerical simulations, we predict polaritonic chiral edge states residing inside the topological gaps. Our Letter paves the way for the further study of strong exciton-photon interaction in nanophotonic structures and for exploring polaritonic topological phases and their practical applications in polaritonic devices.

Fluorescent porous silicon biological probes with high quantum efficiency and stability.

We demonstrate porous silicon biological probes as a stable and non-toxic alternative to organic dyes or cadmium-containing quantum dots for imaging and sensing applications. The fluorescent silicon quantum dots which are embedded on the porous silicon surface are passivated with carboxyl-terminated ligands through stable Si-C covalent bonds. The porous silicon bio-probes have shown photoluminescence quantum yield around 50% under near-UV excitation, with high photochemical and thermal stability. The bio-probes can be efficiently conjugated with antibodies, which is confirmed by a standard enzyme-linked immunosorbent assay (ELISA) method.

Non-volatile electrical polarization switching via domain wall release in 3R-MoS2 bilayer.

Understanding the nature of sliding ferroelectricity is of fundamental importance for the discovery and application of two-dimensional ferroelectric materials. In this work, we investigate the phenomenon of switchable polarization in a bilayer MoS2 with natural rhombohedral stacking, where the spontaneous polarization is coupled with excitonic effects through asymmetric interlayer coupling. Using optical spectroscopy and imaging techniques, we observe how a released domain wall switches the polarization of a large single domain. Our results highlight the importance of domain walls in the polarization switching of non-twisted rhombohedral transition metal dichalcogenides and open new opportunities for the non-volatile control of their optical response.

Efficient Deep Reinforcement Learning With Imitative Expert Priors for Autonomous Driving.

Deep reinforcement learning (DRL) is a promising way to achieve human-like autonomous driving. However, the low sample efficiency and difficulty of designing reward functions for DRL would hinder its applications in practice. In light of this, this article proposes a novel framework to incorporate human prior knowledge in DRL, in order to improve the sample efficiency and save the effort of designing sophisticated reward functions. Our framework consists of three ingredients, namely, expert demonstration, policy derivation, and RL. In the expert demonstration step, a human expert demonstrates their execution of the task, and their behaviors are stored as state-action pairs. In the policy derivation step, the imitative expert policy is derived using behavioral cloning and uncertainty estimation relying on the demonstration data. In the RL step, the imitative expert policy is utilized to guide the learning of the DRL agent by regularizing the KL divergence between the DRL agent's policy and the imitative expert policy. To validate the proposed method in autonomous driving applications, two simulated urban driving scenarios (unprotected left turn and roundabout) are designed. The strengths of our proposed method are manifested by the training results as our method can not only achieve the best performance but also significantly improve the sample efficiency in comparison with the baseline algorithms (particularly 60% improvement compared with soft actor-critic). In testing conditions, the agent trained by our method obtains the highest success rate and shows diverse and human-like driving behaviors as demonstrated by the human expert. We also find that using the imitative expert policy trained with the ensemble method that estimates both policy and model uncertainties, as well as increasing the training sample size, can result in better training and testing performance, especially for more difficult tasks. As a result, the proposed method has shown its potential to facilitate the applications of DRL-enabled human-like autonomous driving systems in practice. The code and supplementary videos are also provided. [https://mczhi.github.io/Expert-Prior-RL/].

Differentiable Integrated Motion Prediction and Planning With Learnable Cost Function for Autonomous Driving.

Predicting the future states of surrounding traffic participants and planning a safe, smooth, and socially compliant trajectory accordingly are crucial for autonomous vehicles (AVs). There are two major issues with the current autonomous driving system: the prediction module is often separated from the planning module, and the cost function for planning is hard to specify and tune. To tackle these issues, we propose a differentiable integrated prediction and planning (DIPP) framework that can also learn the cost function from data. Specifically, our framework uses a differentiable nonlinear optimizer as the motion planner, which takes as input the predicted trajectories of surrounding agents given by the neural network and optimizes the trajectory for the AV, enabling all operations to be differentiable, including the cost function weights. The proposed framework is trained on a large-scale real-world driving dataset to imitate human driving trajectories in the entire driving scene and validated in both open-loop and closed-loop manners. The open-loop testing results reveal that the proposed method outperforms the baseline methods across a variety of metrics and delivers planning-centric prediction results, allowing the planning module to output trajectories close to those of human drivers. In closed-loop testing, the proposed method outperforms various baseline methods, showing the ability to handle complex urban driving scenarios and robustness against the distributional shift. Importantly, we find that joint training of planning and prediction modules achieves better performance than planning with a separate trained prediction module in both open-loop and closed-loop tests. Moreover, the ablation study indicates that the learnable components in the framework are essential to ensure planning stability and performance. Code and Supplementary Videos are available at https://mczhi.github.io/DIPP/.

Human-Guided Reinforcement Learning With Sim-to-Real Transfer for Autonomous Navigation.

Reinforcement learning (RL) is a promising approach in unmanned ground vehicles (UGVs) applications, but limited computing resource makes it challenging to deploy a well-behaved RL strategy with sophisticated neural networks. Meanwhile, the training of RL on navigation tasks is difficult, which requires a carefully-designed reward function and a large number of interactions, yet RL navigation can still fail due to many corner cases. This shows the limited intelligence of current RL methods, thereby prompting us to rethink combining RL with human intelligence. In this paper, a human-guided RL framework is proposed to improve RL performance both during learning in the simulator and deployment in the real world. The framework allows humans to intervene in RL's control progress and provide demonstrations as needed, thereby improving RL's capabilities. An innovative human-guided RL algorithm is proposed that utilizes a series of mechanisms to improve the effectiveness of human guidance, including human-guided learning objective, prioritized human experience replay, and human intervention-based reward shaping. Our RL method is trained in simulation and then transferred to the real world, and we develop a denoised representation for domain adaptation to mitigate the simulation-to-real gap. Our method is validated through simulations and real-world experiments to navigate UGVs in diverse and dynamic environments based only on tiny neural networks and image inputs. Our method performs better in goal-reaching and safety than existing learning- and model-based navigation approaches and is robust to changes in input features and ego kinetics. Furthermore, our method allows small-scale human demonstrations to be used to improve the trained RL agent and learn expected behaviors online.

Sampling Efficient Deep Reinforcement Learning Through Preference-Guided Stochastic Exploration.

Stochastic exploration is the key to the success of the deep Q -network (DQN) algorithm. However, most existing stochastic exploration approaches either explore actions heuristically regardless of their Q values or couple the sampling with Q values, which inevitably introduce bias into the learning process. In this article, we propose a novel preference-guided ϵ -greedy exploration algorithm that can efficiently facilitate exploration for DQN without introducing additional bias. Specifically, we design a dual architecture consisting of two branches, one of which is a copy of DQN, namely, the Q branch. The other branch, which we call the preference branch, learns the action preference that the DQN implicitly follows. We theoretically prove that the policy improvement theorem holds for the preference-guided ϵ -greedy policy and experimentally show that the inferred action preference distribution aligns with the landscape of corresponding Q values. Intuitively, the preference-guided ϵ -greedy exploration motivates the DQN agent to take diverse actions, so that actions with larger Q values can be sampled more frequently, and those with smaller Q values still have a chance to be explored, thus encouraging the exploration. We comprehensively evaluate the proposed method by benchmarking it with well-known DQN variants in nine different environments. Extensive results confirm the superiority of our proposed method in terms of performance and convergence speed.

Transmission-grating-photomasked transient spin grating and its application to measurement of electron-spin ambipolar diffusion in (110) GaAs quantum wells.

A circular dichromatic transient absorption difference spectroscopy of transmission-grating-photomasked transient spin grating is developed and formularized. It is very simple in experimental setup and operation, and has high detection sensitivity. It is applied to measure spin diffusion dynamics and excited electron density dependence of spin ambipolar diffusion coefficient in (110) GaAs quantum wells. It is found that the spin ambipolar diffusion coefficient of (110) and (001) GaAs quantum wells is close to each other, but has an opposite dependence tendency on excited electron density. This spectroscopy is expected to have extensive applicability in the measurement of spin transport.

Prioritized Experience-Based Reinforcement Learning With Human Guidance for Autonomous Driving.

Reinforcement learning (RL) requires skillful definition and remarkable computational efforts to solve optimization and control problems, which could impair its prospect. Introducing human guidance into RL is a promising way to improve learning performance. In this article, a comprehensive human guidance-based RL framework is established. A novel prioritized experience replay mechanism that adapts to human guidance in the RL process is proposed to boost the efficiency and performance of the RL algorithm. To relieve the heavy workload on human participants, a behavior model is established based on an incremental online learning method to mimic human actions. We design two challenging autonomous driving tasks for evaluating the proposed algorithm. Experiments are conducted to access the training and testing performance and learning mechanism of the proposed algorithm. Comparative results against the state-of-the-art methods suggest the advantages of our algorithm in terms of learning efficiency, performance, and robustness.

Ultrafast response of spontaneous photovoltaic effect in 3R-MoS2-based heterostructures.

Rhombohedrally stacked MoS2 has been shown to exhibit spontaneous polarization down to the bilayer limit and can sustain a strong depolarization field when sandwiched between graphene. Such a field gives rise to a spontaneous photovoltaic effect without needing any p-n junction. In this work, we show that the photovoltaic effect has an external quantum efficiency of 10% for devices with only two atomic layers of MoS2 at low temperatures, and identify a picosecond-fast photocurrent response, which translates to an intrinsic device bandwidth at ∼100-GHz level. To this end, we have developed a nondegenerate pump-probe photocurrent spectroscopy technique to deconvolute the thermal and charge-transfer processes, thus successfully revealing the multicomponent nature of the photocurrent dynamics. The fast component approaches the limit of the charge-transfer speed at the graphene-MoS2 interface. The remarkable efficiency and ultrafast photoresponse in the graphene-3R-MoS2 devices support the use of ferroelectric van der Waals materials for future high-performance optoelectronic applications.

Ultrathin (<1 µm) Substrate-Free Flexible Photodetector on Quantum Dot-Nanocellulose Paper.

Conventional approaches to flexible optoelectronic devices typically require depositing the active materials on external substrates. This is mostly due to the weak bonding between individual molecules or nanocrystals in the active materials, which prevents sustaining a freestanding thin film. Herein we demonstrate an ultrathin freestanding ZnO quantum dot (QD) active layer with nanocellulose structuring, and its corresponding device fabrication method to achieve substrate-free flexible optoelectronic devices. The ultrathin ZnO QD-nanocellulose composite is obtained by hydrogel transfer printing and solvent-exchange processes to overcome the water capillary force which is detrimental to achieving freestanding thin films. We achieved an active nanocellulose paper with ~550 nm thickness, and >91% transparency in the visible wavelength range. The film retains the photoconductive and photoluminescent properties of ZnO QDs and is applied towards substrate-free Schottky photodetector applications. The device has an overall thickness of ~670 nm, which is the thinnest freestanding optoelectronic device to date, to the best of our knowledge, and functions as a self-powered visible-blind ultraviolet photodetector. This platform can be readily applied to other nano materials as well as other optoelectronic device applications.

Photonic Crystal Optical Tweezers with High Efficiency for Live Biological Samples and Viability Characterization.

We propose and demonstrate a new optical trapping method for single cells that utilizes modulated light fields to trap a wide array of cell types, including mammalian, yeast, and Escherichia coli cells, on the surface of a two-dimensional photonic crystal. This method is capable of reducing the required light intensity, and thus minimizing the photothermal damage to living cells, thereby extending cell viability in optical trapping and cell manipulation applications. To this end, a thorough characterization of cell viability in optical trapping environments was performed. This study also demonstrates the technique using spatial light modulation in patterned manipulation of live cell arrays over a broad area.