Our work focuses on the development of a learnable neural representation of human pose for advanced AI assisted animation tooling. Specifically, we tackle the problem of constructing a full static human pose based on sparse and variable user inputs (e.g. locations and/or orientations of a subset of body joints). To solve this problem, we propose a novel neural architecture that combines residual connections with prototype encoding of a partially specified pose to create a new complete pose from the learned latent space. We show that our architecture outperforms a baseline based on Transformer, both in terms of accuracy and computational efficiency. Additionally, we develop a user interface to integrate our neural model in Unity, a real-time 3D development platform. Furthermore, we introduce two new datasets representing the static human pose modeling problem, based on high-quality human motion capture data, which will be released publicly along with model code.
We consider a generalization of rational inattention problems by measuring costs of information through the information radius (Sibson, 1969; Verd\'u, 2015) of statistical experiments. We introduce a notion of attention elasticity measuring the sensitivity of attention strategies with respect to changes in incentives. We show how the introduced class of cost functions controls attention elasticities while the Shannon model restricts attention elasticity to be unity. We explore further differences and similarities relative to the Shannon model in relation to invariance, posterior separability, consideration sets, and the ability to learn events with certainty. Lastly, we provide an efficient alternating minimization method -- analogous to the Blahut-Arimoto algorithm -- to obtain optimal attention strategies.
Digital Twin, as an emerging technology related to Cyber-Physical Systems (CPS) and Internet of Things (IoT), has attracted increasing attentions during the past decade. Conceptually, a Digital Twin is a digital replica of a physical entity in the real world, and this technology is leveraged in this study to design a cooperative driving system at non-signalized intersections, allowing connected vehicles to cooperate with each other to cross intersections without any full stops. Within the proposed Digital Twin framework, we developed an enhanced first-in-first-out (FIFO) slot reservation algorithm to schedule the sequence of crossing vehicles, a consensus motion control algorithm to calculate vehicles' referenced longitudinal motion, and a model-based motion estimation algorithm to tackle communication delay and packet loss. Additionally, an augmented reality (AR) human-machine-interface (HMI) is designed to provide the guidance to drivers to cooperate with other connected vehicles. Agent-based modeling and simulation of the proposed system is conducted in Unity game engine based on a real-world map in San Francisco, and the human-in-the-loop (HITL) simulation results prove the benefits of the proposed algorithms with 20% reduction in travel time and 23.7% reduction in energy consumption, respectively, when compared with traditional signalized intersections.
Intelligent reflective surface (IRS) technology is emerging as a promising performance enhancement technique for next-generation wireless networks. Hence, we investigate the physical layer security of the downlink in IRS-aided non-orthogonal multiple access networks in the presence of an eavesdropper, where an IRS is deployed for enhancing the quality by assisting the cell-edge user to communicate with the base station. To characterize the network's performance, the expected value of the new channel statistics is derived for the reflected links in the case of Nakagami-m fading. Furthermore, the performance of the proposed network is evaluated both in terms of the secrecy outage probability (SOP) and the average secrecy capacity (ASC). The closed-form expressions of the SOP and the ASC are derived. We also study the impact of various network parameters on the overall performance of the network considered. To obtain further insights, the secrecy diversity orders and the high signal-to-noise ratio slopes are obtained. We finally show that: 1) the expectation of the channel gain in the reflected links is determined both by the number of IRSs and by the Nakagami-m fading parameters; 2) The SOP of both receiver 1 and receiver 2 becomes unity, when the number of IRSs is sufficiently high; 3) The secrecy diversity orders are affected both by the number of IRSs and by the Nakagami-m fading parameters, whereas the high-SNR slopes are not affected by these parameters. Our Monte-Carlo simulations perfectly demonstrate the analytical results.
Partition of unity methods (PUMs) on graphs are simple and highly adaptive auxiliary tools for graph signal processing. Based on a greedy-type metric clustering and augmentation scheme, we show how a partition of unity can be generated in an efficient way on graphs. We investigate how PUMs can be combined with a local graph basis function (GBF) approximation method in order to obtain low-cost global interpolation or classification schemes. From a theoretical point of view, we study necessary prerequisites for the partition of unity such that global error estimates of the PUM follow from corresponding local ones. Finally, properties of the PUM as cost-efficiency and approximation accuracy are investigated numerically.
Deep generative frameworks including GANs and normalizing flow models have proven successful at filling in missing values in partially observed data samples by effectively learning -- either explicitly or implicitly -- complex, high-dimensional statistical distributions. In tasks where the data available for learning is only partially observed, however, their performance decays monotonically as a function of the data missingness rate. In high missing data rate regimes (e.g., 60% and above), it has been observed that state-of-the-art models tend to break down and produce unrealistic and/or semantically inaccurate data. We propose a novel framework to facilitate the learning of data distributions in high paucity scenarios that is inspired by traditional formulations of solutions to ill-posed problems. The proposed framework naturally stems from posing the process of learning from incomplete data as a joint optimization task of the parameters of the model being learned and the missing data values. The method involves enforcing a prior regularization term that seamlessly integrates with objectives used to train explicit and tractable deep generative frameworks such as deep normalizing flow models. We demonstrate via extensive experimental validation that the proposed framework outperforms competing techniques, particularly as the rate of data paucity approaches unity.
This paper is concerned with the design of intermittent non-pharmaceutical strategies to mitigate the spread of the COVID-19 epidemic exploiting network epidemiological models. Specifically, by studying a variational equation for the dynamics of the infected in a network model of the epidemic spread, we derive, using contractivity arguments, a condition that can be used to guarantee that, in epidemiological terms, the effective reproduction number is less than unity. This condition has three advantages: (i) it is easily computable; (ii) it is directly related to the model parameters; (iii) it can be used to enforce a scalability condition that prohibits the amplification of disturbances within the network system. We then include satisfaction of such a condition as a constraint in a Model Predictive Control problem so as to mitigate (or suppress) the spread of the epidemic while minimizing the economic impact of the interventions. A data-driven model of Italy as a network of three macro-regions (North, Center, and South), whose parameters are identified from real data, is used to illustrate and evaluate the effectiveness of the proposed control strategy.
The dynamics of epidemic spreading is often reduced to the single control parameter $R_0$, whose value, above or below unity, determines the state of the contagion. If, however, the pathogen evolves as it spreads, $R_0$ may change over time, potentially leading to a mutation-driven spread, in which an initially sub-pandemic pathogen undergoes a breakthrough mutation. To predict the boundaries of this pandemic phase, we introduce here a modeling framework to couple the network spreading patterns with the intra-host evolutionary dynamics. For many pathogens these two processes, intra- and inter-host, are driven by different selection forces. And yet here we show that even in the extreme case when these two forces are mutually independent, mutations can still fundamentally alter the pandemic phase-diagram, whose transitions are now shaped, not just by $R_0$, but also by the balance between the epidemic and the evolutionary timescales. If mutations are too slow, the pathogen prevalence decays prior to the appearance of a critical mutation. On the other hand, if mutations are too rapid, the pathogen evolution becomes volatile and, once again, it fails to spread. Between these two extremes, however, we identify a broad range of conditions in which an initially sub-pandemic pathogen can break through to gain widespread prevalence.
Along with the development of modern computing technology and social sciences, both theoretical research and practical applications of social computing have been continuously extended. In particular with the boom of artificial intelligence (AI), social computing is significantly influenced by AI. However, the conventional technologies of AI have drawbacks in dealing with more complicated and dynamic problems. Such deficiency can be rectified by hybrid human-artificial intelligence (H-AI) which integrates both human intelligence and AI into one unity, forming a new enhanced intelligence. H-AI in dealing with social problems shows the advantages that AI can not surpass. This paper firstly introduces the concept of H-AI. AI is the intelligence in the transition stage of H-AI, so the latest research progresses of AI in social computing are reviewed. Secondly, it summarizes typical challenges faced by AI in social computing, and makes it possible to introduce H-AI to solve these challenges. Finally, the paper proposes a holistic framework of social computing combining with H-AI, which consists of four layers: object layer, base layer, analysis layer, and application layer. It represents H-AI has significant advantages over AI in solving social problems.
This paper describes an adaptive method in continuous time for the estimation of external fields by a team of $N$ agents. The agents $i$ each explore subdomains $\Omega^i$ of a bounded subset of interest $\Omega\subset X := \mathbb{R}^d$. Ideal adaptive estimates $\hat{g}^i_t$ are derived for each agent from a distributed parameter system (DPS) that takes values in the scalar-valued reproducing kernel Hilbert space $H_X$ of functions over $X$. Approximations of the evolution of the ideal local estimate $\hat{g}^i_t$ of agent $i$ is constructed solely using observations made by agent $i$ on a fine time scale. Since the local estimates on the fine time scale are constructed independently for each agent, we say that the method is strictly decentralized. On a coarse time scale, the individual local estimates $\hat{g}^i_t$ are fused via the expression $\hat{g}_t:=\sum_{i=1}^N\Psi^i \hat{g}^i_t$ that uses a partition of unity $\{\Psi^i\}_{1\leq i\leq N}$ subordinate to the cover $\{\Omega^i\}_{i=1,\ldots,N}$ of $\Omega$. Realizable algorithms are obtained by constructing finite dimensional approximations of the DPS in terms of scattered bases defined by each agent from samples along their trajectories. Rates of convergence of the error in the finite dimensional approximations are derived in terms of the fill distance of the samples that define the scattered centers in each subdomain. The qualitative performance of the convergence rates for the decentralized estimation method is illustrated via numerical simulations.
Using a mean-field theory of signal propagation, we analyze the evolution of correlations between two signals propagating forward through a deep ReLU network with correlated weights. Signals become highly correlated in deep ReLU networks with uncorrelated weights. We show that ReLU networks with anti-correlated weights can avoid this fate and have a chaotic phase where the signal correlations saturate below unity. Consistent with this analysis, we find that networks initialized with anti-correlated weights can train faster (in a teacher-student setting) by taking advantage of the increased expressivity in the chaotic phase. Combining this with a previously proposed strategy of using an asymmetric initialization to reduce dead node probability, we propose an initialization scheme that allows faster training and learning than the best-known initializations.
Automatic image aesthetics assessment is a computer vision problem dealing with categorizing images into different aesthetic levels. The categorization is usually done by analyzing an input image and computing some measure of the degree to which the image adheres to the fundamental principles of photography such as balance, rhythm, harmony, contrast, unity, look, feel, tone and texture. Due to its diverse applications in many areas, automatic image aesthetic assessment has gained significant research attention in recent years. This article presents a review of the contemporary automatic image aesthetics assessment techniques. Many traditional hand-crafted and deep learning-based approaches are reviewed, and critical problem aspects are discussed, including why some features or models perform better than others and the limitations. A comparison of the quantitative results of different methods is also provided.
Deep neural networks unlocked a vast range of new applications by solving tasks of which many were previously deemed as reserved to higher human intelligence. One of the developments enabling this success was a boost in computing power provided by special purpose hardware, such as graphic or tensor processing units. However, these do not leverage fundamental features of neural networks like parallelism and analog state variables. Instead, they emulate neural networks relying on binary computing, which results in unsustainable energy consumption and comparatively low speed. Fully parallel and analogue hardware promises to overcome these challenges, yet the impact of analogue neuron noise and its propagation, i.e. accumulation, threatens rendering such approaches inept. Here, we determine for the first time the propagation of noise in deep neural networks comprising noisy nonlinear neurons in trained fully connected layers. We study additive and multiplicative as well as correlated and uncorrelated noise, and develop analytical methods that predict the noise level in any layer of symmetric deep neural networks or deep neural networks trained with back propagation. We find that noise accumulation is generally bound, and adding additional network layers does not worsen the signal to noise ratio beyond a limit. Most importantly, noise accumulation can be suppressed entirely when neuron activation functions have a slope smaller than unity. We therefore developed the framework for noise in fully connected deep neural networks implemented in analog systems, and identify criteria allowing engineers to design noise-resilient novel neural network hardware.
The problem of mixed static and dynamic obstacle avoidance is essential for path planning in highly dynamic environment. However, the paths formed by grid edges can be longer than the true shortest paths in the terrain since their headings are artificially constrained. Existing methods can hardly deal with dynamic obstacles. To address this problem, we propose a new algorithm combining Model Predictive Control (MPC) with Deep Deterministic Policy Gradient (DDPG). Firstly, we apply the MPC algorithm to predict the trajectory of dynamic obstacles. Secondly, the DDPG with continuous action space is designed to provide learning and autonomous decision-making capability for robots. Finally, we introduce the idea of the Artificial Potential Field to set the reward function to improve convergence speed and accuracy. We employ Unity 3D to perform simulation experiments in highly uncertain environment such as aircraft carrier decks and squares. The results show that our method has made great improvement on accuracy by 7%-30% compared with the other methods, and on the length of the path and turning angle by reducing 100 units and 400-450 degrees compared with DQN (Deep Q Network), respectively.
There has been rapidly growing interest in meta-learning as a method for increasing the flexibility and sample efficiency of reinforcement learning. One problem in this area of research, however, has been a scarcity of adequate benchmark tasks. In general, the structure underlying past benchmarks has either been too simple to be inherently interesting, or too ill-defined to support principled analysis. In the present work, we introduce a new benchmark for meta-RL research, emphasizing transparency and potential for in-depth analysis as well as structural richness. Alchemy is a 3D video game, implemented in Unity, which involves a latent causal structure that is resampled procedurally from episode to episode, affording structure learning, online inference, hypothesis testing and action sequencing based on abstract domain knowledge. We evaluate a pair of powerful RL agents on Alchemy and present an in-depth analysis of one of these agents. Results clearly indicate a frank and specific failure of meta-learning, providing validation for Alchemy as a challenging benchmark for meta-RL. Concurrent with this report, we are releasing Alchemy as public resource, together with a suite of analysis tools and sample agent trajectories.
Generative neural networks have a well recognized ability to estimate underlying manifold structure of high dimensional data. However, if a single latent space is used, it is not possible to faithfully represent a manifold with topology different from Euclidean space. In this work we define the general class of Atlas Generative Models (AGMs), models with hybrid discrete-continuous latent space that estimate an atlas on the underlying data manifold together with a partition of unity on the data space. We identify existing examples of models from various popular generative paradigms that fit into this class. Due to the atlas interpretation, ideas from non-linear latent space analysis and statistics, e.g. geodesic interpolation, which has previously only been investigated for models with simply connected latent spaces, may be extended to the entire class of AGMs in a natural way. We exemplify this by generalizing an algorithm for graph based geodesic interpolation to the setting of AGMs, and verify its performance experimentally.
Approximation theorists have established best-in-class optimal approximation rates of deep neural networks by utilizing their ability to simultaneously emulate partitions of unity and monomials. Motivated by this, we propose partition of unity networks (POUnets) which incorporate these elements directly into the architecture. Classification architectures of the type used to learn probability measures are used to build a meshfree partition of space, while polynomial spaces with learnable coefficients are associated to each partition. The resulting hp-element-like approximation allows use of a fast least-squares optimizer, and the resulting architecture size need not scale exponentially with spatial dimension, breaking the curse of dimensionality. An abstract approximation result establishes desirable properties to guide network design. Numerical results for two choices of architecture demonstrate that POUnets yield hp-convergence for smooth functions and consistently outperform MLPs for piecewise polynomial functions with large numbers of discontinuities.
This paper proposes an agent-based simulation of a presidential election, inspired by the French 2017 presidential election. The simulation is based on data extracted from polls, media coverage, and Twitter. The main contribution is to consider the impact of scandals and media bashing on the result of the election. In particular, it is shown that scandals can lead to higher abstention at the election, as voters have no relevant candidate left to vote for. The simulation is implemented in Unity 3D and is available to play online.
Traditional and deep learning-based fusion methods generated the intermediate decision map to obtain the fusion image through a series of post-processing procedures. However, the fusion results generated by these methods are easy to lose some source image details or results in artifacts. Inspired by the image reconstruction techniques based on deep learning, we propose a multi-focus image fusion network framework without any post-processing to solve these problems in the end-to-end and supervised learning way. To sufficiently train the fusion model, we have generated a large-scale multi-focus image dataset with ground-truth fusion images. What's more, to obtain a more informative fusion image, we further designed a novel fusion strategy based on unity fusion attention, which is composed of a channel attention module and a spatial attention module. Specifically, the proposed fusion approach mainly comprises three key components: feature extraction, feature fusion and image reconstruction. We firstly utilize seven convolutional blocks to extract the image features from source images. Then, the extracted convolutional features are fused by the proposed fusion strategy in the feature fusion layer. Finally, the fused image features are reconstructed by four convolutional blocks. Experimental results demonstrate that the proposed approach for multi-focus image fusion achieves remarkable fusion performance compared to 19 state-of-the-art fusion methods.
Simulations, along with other similar applications like virtual worlds and video games, require computational models of intelligence that generate realistic and credible behavior for the participating synthetic characters. Cognitive architectures, which are models of the fixed structure underlying intelligent behavior in both natural and artificial systems, provide a conceptually valid common basis, as evidenced by the current efforts towards a standard model of the mind, to generate human-like intelligent behavior for these synthetic characters. Sigma is a cognitive architecture and system that strives to combine what has been learned from four decades of independent work on symbolic cognitive architectures, probabilistic graphical models, and more recently neural models, under its graphical architecture hypothesis. Sigma leverages an extended form of factor graphs towards a uniform grand unification of not only traditional cognitive capabilities but also key non-cognitive aspects, creating unique opportunities for the construction of new kinds of cognitive models that possess a Theory-of-Mind and that are perceptual, autonomous, interactive, affective, and adaptive. In this paper, we will introduce Sigma along with its diverse capabilities and then use three distinct proof-of-concept Sigma models to highlight combinations of these capabilities: (1) Distributional reinforcement learning models in; (2) A pair of adaptive and interactive agent models that demonstrate rule-based, probabilistic, and social reasoning; and (3) A knowledge-free exploration model in which an agent leverages only architectural appraisal variables, namely attention and curiosity, to locate an item while building up a map in a Unity environment.