This paper introduces a new concept. We intend to give life to a software agent. A software agent is a computer program that acts on a user's behalf. We put a DNA inside the agent. DNA is a simple text, a whole roadmap of a network of agents or a system with details. A Dynamic Numerical Abstract of a multiagent system. It is also a reproductive part for an \emph{agent} that makes the agent take actions and decide independently and reproduce coworkers. By defining different DNA structures, one can establish new agents and different nets for different usages. We initiate such thinking as \emph{DNA programming}. This strategy leads to a new field of programming. This type of programming can help us manage large systems with various elements with an incredibly organized customizable structure. An agent can reproduce another agent. We put one or a few agents around a given network, and the agents will reproduce themselves till they can reach others and pervade the whole network. An agent's position or other environmental or geographical characteristics make it possible for an agent to know its active set of \emph{genes} on its DNA. The active set of genes specifies its duties. There is a database that includes a list of functions s.t. each one is an implementation of what a \emph{gene} represents. To utilize a decentralized database, we may use a blockchain-based structure. This design can adapt to a system that manages many static and dynamic networks. This network could be a distributed system, a decentralized system, a telecommunication network such as a 5G monitoring system, an IoT management system, or even an energy management system. The final system is the combination of all the agents and the overlay net that connects the agents. We denote the final net as the \emph{body} of the system.

Recurrent neural networks are effective models to process sequences. However, they are unable to learn long-term dependencies because of their inherent sequential nature. As a solution, Vaswani et al. introduced the Transformer, a model solely based on the attention mechanism that is able to relate any two positions of the input sequence, hence modelling arbitrary long dependencies. The Transformer has improved the state-of-the-art across numerous sequence modelling tasks. However, its effectiveness comes at the expense of a quadratic computational and memory complexity with respect to the sequence length, hindering its adoption. Fortunately, the deep learning community has always been interested in improving the models' efficiency, leading to a plethora of solutions such as parameter sharing, pruning, mixed-precision, and knowledge distillation. Recently, researchers have directly addressed the Transformer's limitation by designing lower-complexity alternatives such as the Longformer, Reformer, Linformer, and Performer. However, due to the wide range of solutions, it has become challenging for researchers and practitioners to determine which methods to apply in practice in order to meet the desired trade-off between capacity, computation, and memory. This survey addresses this issue by investigating popular approaches to make Transformers faster and lighter and by providing a comprehensive explanation of the methods' strengths, limitations, and underlying assumptions.

One of the most difficult parts of motion planning in configuration space is ensuring a trajectory does not collide with task-space obstacles in the environment. Generating regions that are convex and collision free in configuration space can separate the computational burden of collision checking from motion planning. To that end, we propose an extension to IRIS (Iterative Regional Inflation by Semidefinite programming) [5] that allows it to operate in configuration space. Our algorithm, IRIS-NP (Iterative Regional Inflation by Semidefinite & Nonlinear Programming), uses nonlinear optimization to add the separating hyperplanes, enabling support for more general nonlinear constraints. Developed in parallel to Amice et al. [1], IRIS-NP trades rigorous certification that regions are collision free for probabilistic certification and the benefit of faster region generation in the configuration-space coordinates. IRIS-NP also provides a solid initialization to C-IRIS to reduce the number of iterations required for certification. We demonstrate that IRIS-NP can scale to a dual-arm manipulator and can handle additional nonlinear constraints using the same machinery. Finally, we show ablations of elements of our implementation to demonstrate their importance.

Point cloud data have been widely explored due to its superior accuracy and robustness under various adverse situations. Meanwhile, deep neural networks (DNNs) have achieved very impressive success in various applications such as surveillance and autonomous driving. The convergence of point cloud and DNNs has led to many deep point cloud models, largely trained under the supervision of large-scale and densely-labelled point cloud data. Unsupervised point cloud representation learning, which aims to learn general and useful point cloud representations from unlabelled point cloud data, has recently attracted increasing attention due to the constraint in large-scale point cloud labelling. This paper provides a comprehensive review of unsupervised point cloud representation learning using DNNs. It first describes the motivation, general pipelines as well as terminologies of the recent studies. Relevant background including widely adopted point cloud datasets and DNN architectures is then briefly presented. This is followed by an extensive discussion of existing unsupervised point cloud representation learning methods according to their technical approaches. We also quantitatively benchmark and discuss the reviewed methods over multiple widely adopted point cloud datasets. Finally, we share our humble opinion about several challenges and problems that could be pursued in future research in unsupervised point cloud representation learning. A project associated with this survey has been built at https://github.com/xiaoaoran/3d_url_survey.

Energetic materials (EM) cover a wide spectrum of propellants, pyrotechnics, and explosives and are key components in military applications for propulsion and munition systems and in civilian applications such as construction and mining [1]. Heterogenous/composite EMs have complex microstructures which significantly influence--along with chemistry--the property and performance of these materials [2-8]. There is increasing research interest in controlling the microstructure of EM, to engineer their properties and performance for targeted functional specificity [9-10]. EMs are typically solid-solid composites of organic energetic crystals (commonly CHNO compounds), inclusions (i.e., metals, nanoparticles), and plastic binders. The CHNO materials are commonly categorized based on how sensitive they are to an external load/mechanical insult. They can range f rom'insensitive' (such as TATB - based EMs [11]) to'highly sensitive' (PETN-based EMs [12-13]) with others such as HMX, CL-20, and RDX ranging in between [14]. The sensitivity is closely connected with the molecular structure of these species of EMs within the CHNO family. However, when they are formed into propellants and explosives, the sensitivity is also impacted by the physical structure, composition, and formulation of the material mixtures, as reviewed by Handley et al. [1]. In other words, the design of a mixture and its microstructure can define the overall properties and performance characteristics of formed EM, thus opening the possibility of systematic methods to engineer materials by their design.

When training neural networks for classification tasks with backpropagation, parameters are updated on every trial, even if the sample is classified correctly. In contrast, humans concentrate their learning effort on errors. Inspired by human learning, we introduce lazy learning, which only learns on incorrect samples. Lazy learning can be implemented in a few lines of code and requires no hyperparameter tuning. Lazy learning achieves state-of-the-art performance and is particularly suited when datasets are large. For instance, it reaches 99.2% test accuracy on Extended MNIST using a single-layer MLP, and does so 7.6 faster than a matched backprop network. Recent progress in machine learning has been partly attributed to the use of large data sets [LeCun et al., 2015]. Even already large datasets are often augmented to further boost performance. However, repeatedly cycling over large datasets and adjusting the parameters is time and energy consuming. In classification tasks, backprop typically prescribes synaptic updates regardless of whether the classification was correct or incorrect; updating the network to be correct if it was wrong, but also updating to be more correct if it was right.

Automated co-design of machine learning models and evaluation hardware is critical for efficiently deploying such models at scale. Despite the state-of-the-art performance of transformer models, they are not yet ready for execution on resource-constrained hardware platforms. High memory requirements and low parallelizability of the transformer architecture exacerbate this problem. Recently-proposed accelerators attempt to optimize the throughput and energy consumption of transformer models. However, such works are either limited to a one-sided search of the model architecture or a restricted set of off-the-shelf devices. Furthermore, previous works only accelerate model inference and not training, which incurs substantially higher memory and compute resources, making the problem even more challenging. To address these limitations, this work proposes a dynamic training framework, called DynaProp, that speeds up the training process and reduces memory consumption. DynaProp is a low-overhead pruning method that prunes activations and gradients at runtime. To effectively execute this method on hardware for a diverse set of transformer architectures, we propose ELECTOR, a framework that simulates transformer inference and training on a design space of accelerators. We use this simulator in conjunction with the proposed co-design technique, called TransCODE, to obtain the best-performing models with high accuracy on the given task and minimize latency, energy consumption, and chip area. The obtained transformer-accelerator pair achieves 0.3% higher accuracy than the state-of-the-art pair while incurring 5.2$\times$ lower latency and 3.0$\times$ lower energy consumption.

The synthesis of tactile sensing with compliance is essential to many fields, from agricultural usages like fruit picking, to sustainability practices such as sorting recycling, to the creation of safe home-care robots for the elderly to age with dignity. From tactile sensing, we can discern material properties, recognize textures, and determine softness, while with compliance, we are able to securely and safely interact with the objects and the environment around us. These two abilities can culminate into a useful soft robotic gripper, such as the original GelSight Fin Ray, which is able to grasp a large variety of different objects and also perform a simple household manipulation task: wine glass reorientation. Although the original GelSight Fin Ray solves the problem of interfacing a generally rigid, high-resolution sensor with a soft, compliant structure, we can improve the robustness of the sensor and implement techniques that make such camera-based tactile sensors applicable to a wider variety of soft robot designs. We first integrate flexible mirrors and incorporate the rigid electronic components into the base of the gripper, which greatly improves the compliance of the Fin Ray structure. Then, we synthesize a flexible and high-elongation silicone adhesive-based fluorescent paint, which can provide good quality 2D tactile localization results for our sensor. Finally, we incorporate all of these techniques into a new design: the Baby Fin Ray, which we use to dig through clutter, and perform successful classification of nuts in their shells. The supplementary video can be found here: https://youtu.be/_oD_QFtYTPM

The ability to forecast human-environment collisions from egocentric observations is vital to enable collision avoidance in applications such as VR, AR, and wearable assistive robotics. In this work, we introduce the challenging problem of predicting collisions in diverse environments from multi-view egocentric videos captured from body-mounted cameras. Solving this problem requires a generalizable perception system that can classify which human body joints will collide and estimate a collision region heatmap to localize collisions in the environment. To achieve this, we propose a transformer-based model called COPILOT to perform collision prediction and localization simultaneously, which accumulates information across multi-view inputs through a novel 4D space-time-viewpoint attention mechanism. To train our model and enable future research on this task, we develop a synthetic data generation framework that produces egocentric videos of virtual humans moving and colliding within diverse 3D environments. This framework is then used to establish a large-scale dataset consisting of 8.6M egocentric RGBD frames. Extensive experiments show that COPILOT generalizes to unseen synthetic as well as real-world scenes. We further demonstrate COPILOT outputs are useful for downstream collision avoidance through simple closed-loop control. Please visit our project webpage at https://sites.google.com/stanford.edu/copilot.

Moir\'e engineering in atomically thin van der Waals heterostructures creates artificial quantum materials with designer properties. We solve the many-body problem of interacting electrons confined to a moir\'e superlattice potential minimum (the moir\'e atom) using a 2D fermionic neural network. We show that strong Coulomb interactions in combination with the anisotropic moir\'e potential lead to striking ``Wigner molecule" charge density distributions observable with scanning tunneling microscopy.