A deep Koopman operator-based modelling approach for long-term prediction of dynamics with pixel-level measurements

Although previous studies have made some clear leap in learning latent dynamics from high-dimensional representations,the performances in terms of accuracy and inference time of long-term model prediction still need to be improved.In this study,a deep convolutional network based on the Koopman operator(CKNet)is proposed to model non-linear systems with pixel-level measurements for long-term prediction.CKNet adopts an autoencoder network architecture,consisting of an encoder to generate latent states and a linear dynamical model(i.e.,the Koopman operator)which evolves in the latent state space spanned by the encoder.The decoder is used to recover images from latent states.According to a multi-step ahead prediction loss function,the system matrices for approximating the Koopman operator are trained synchronously with the autoencoder in a mini-batch manner.In this manner,gradients can be synchronously transmitted to both the system matrices and the autoencoder to help the encoder self-adaptively tune the latent state space in the training process,and the resulting model is time-invariant in the latent space.Therefore,the proposed CKNet has the advantages of less inference time and high accuracy for long-term prediction.Experiments are per-formed on OpenAI Gym and Mujoco environments,including two and four non-linear forced dynamical systems with continuous action spaces.The experimental results show that CKNet has strong long-term prediction capabilities with sufficient precision.

Unique and complementary suppression of cGAS-STING and RNA sensing- triggered innate immune responses by SARS-CoV-2 proteins

The emergence of SARS-CoV-2 has resulted in the COVID-19 pandemic,leading to millions of infections and hundreds of thousands of human deaths.The efficient replication and population spread of SARS-CoV-2 indicates an effective evasion of human innate immune responses,although the viral proteins responsible for this immune evasion are not clear.In this study,we identified SARS-CoV-2 structural proteins,accessory proteins,and the main viral protease as potent inhibitors of host innate immune responses of distinct pathways.In particular,the main viral protease was a potent inhibitor of both the RLR and cGAS-STING pathways.Viral accessory protein 0RF3a had the unique ability to inhibit STING,but not the RLR response.On the other hand,structural protein N was a unique RLR inhibitor.0RF3a bound STING in a unique fashion and blocked the nuclear accumulation of p65 to inhibit nuclear factor-KB signaling.3CL of SARS-CoV-2 inhibited K63-ubiquitin modification of STING to disrupt the assembly of the STING functional complex and downstream signaling.Diverse vertebrate STINGs,including those from humans,mice,and chickens,could be inhibited by 0RF3a and 3CL of SARS-CoV-2.The existence of more effective innate immune suppressors in pathogenic coronaviruses may allow them to replicate more efficiently in vivo.Since evasion of host innate immune responses is essential for the survival of all viruses,our study provides insights into the design of therapeutic agents against SARS-CoV-2.

Nanocellulose from various biomass wastes:Its preparation and potential usages towards the high value-added products

Biomass waste comes from a wide range of sources,such as forest,agricultural,algae wastes,as well as other relevant industrial by-products.It is an important alternative energy source as well as a unique source for various bioproducts applied in many fields.For the past two decades,how to reuse,recycle and best recover various biomass wastes for high value-added bioproducts has received significant attention,which has not only come from various academia communities but also from many civil and medical industries.To summarize one of the cutting-edge technologies applied with nanocellulose biomaterials,this review focused on various preparation methods and strategies to make nanocellulose from diverse biomass wastes and their potential applications in biomedical areas and other promising new fields.

Recurrent 3D attentional networks for end-to-end active object recognition

Active vision is inherently attention-driven:an agent actively selects views to attend in order to rapidly perform a vision task while improving its internal representation of the scene being observed.Inspired by the recent success of attention-based models in 2D vision tasks based on single RGB images, we address multi-view depth-based active object recognition using an attention mechanism, by use of an end-to-end recurrent 3D attentional network. The architecture takes advantage of a recurrent neural network to store and update an internal representation. Our model,trained with 3D shape datasets, is able to iteratively attend the best views targeting an object of interest for recognizing it. To realize 3D view selection, we derive a 3D spatial transformer network. It is dierentiable,allowing training with backpropagation, and so achieving much faster convergence than the reinforcement learning employed by most existing attention-based models. Experiments show that our method, with only depth input, achieves state-of-the-art next-best-view performance both in terms of time taken and recognition accuracy.