Sum-Product Networks (SPNs) are an alternative to the widely used Neural Networks (NNs) for machine learning. SPNs can not only reason about (un)certainty by qualifying their output with a probability, they also allow fast (tractable) inference by having run-times that are just linear w.r.t. the network size.We present SPNC, the first tool flow for generating fast native code for SPN inference on both CPUs and GPUs,including the use of vectorized/SIMD execution. To this end, we add two SPN-specific dialects to the MLIR framework and discuss their lowering towards the execution targets.We evaluate our approach on two applications, for which we consider performance, scaling to very large SPNs, and compile vs execution-time trade-offs. In this manner, we achieve multiple orders of magnitude in speed-ups over existing SPN support libraries.

Physical simulation is an indispensable component of robotics simulation platforms that serves as the basis for a plethora of research directions. Looking strictly at robotics, the common characteristic of the most popular physics engines, such as ODE, DART, MuJoCo, bullet, SimBody, PhysX or RaiSim, is that they focus on the solution of articulated rigid bodies with collisions and contacts problems, while paying less attention to other physical phenomena. This restriction limits the range of addressable simulation problems, rendering applications such as soft robotics, cloth simulation, simulation of viscoelastic materials, and fluid dynamics, especially surface swimming, infeasible. In this work, we present Gazebo Fluids, an open-source extension of the popular Gazebo robotics simulator that enables the interaction of articulated rigid body dynamics with particle-based fluid and deformable solid simulation. We implement fluid dynamics and highly viscous and elastic material simulation capabilities based on the Smoothed Particle Hydrodynamics method. We demonstrate the practical impact of this extension for previously infeasible application scenarios in a series of experiments, showcasing one of the first self-propelled robot swimming simulations with SPH in a robotics simulator.

Probabilistic models are receiving increasing attention as a complementary alternative to more widespread machine learning approaches, such as neural networks. One particularly interesting class of models are so-called Sum-Product Networks (SPN), which combine the expressiveness of probabilistic models with tractable inference, making them an interesting candidate for use in real-world applications.Yet, as Sum-Product Networks are a young class of machine learning models, the software ecosystem is comparably sparse. In this work, we enhance the ecosystem with a domain-specific compiler that allows to easily and efficiently target CPUs and GPUs for Sum-Product Network inference.Using a real-world application of Sum-Product Networks, a robust speaker identification model, we showcase the performance improvements our compiler can achieve for SPN inference on CPUs and GPUs.

Understanding tumor's evolution under chemotherapy is central in the design of cancer therapy regimens. Drug resistance poses a major obstacle in the battle against most types of cancer and therapy design. Personalized treatments have the potential to offer greater effectiveness and the ability to prevent and circumvent drug resistance. In this study we introduce PERFECTO (Prediction of Extended Response and Growth Functions for Estimating ChemoTherapy Outcomes), a machine learning system capable of extracting the tumor growth function and response under chemotherapy. Exploiting the underlying correlations in the clinical data, the system captures the statistical peculiarities of tumor growth in-vivo without an explicit modeling of tumor microenvironment and expensive clinical investigations. We demonstrate the learning capabilities of PERFECTO in predicting unperturbed tumor growth and chemotherapy tumor growth from multiple clinical breast cancer datasets. We postulate that predictability is the key. Using PERFECTO clinicians will be able to improve treatment plans for patient-specific parameters from individual tumors. Our preliminary experiments on in-vitro, animal and in-vivo datasets, shown that, with a high degree of confidence, PERFECTO is able to estimate treatment effectiveness through an accurate tumor growth response prediction, independent of the breast cancer cell line. This in turn can alleviate the need of ordering extra clinical tests or any extra wait time before treatment initiation.

Modelling surgical size is not inherently meant to replicate the tumor's exact form and proportions, but instead to elucidate the degree of the tissue volume that may be surgically removed in terms of improving patient survival and minimize the risk that subsequent operations will be needed to eliminate all malignant cells entirely. Given the broad range of models of tumor growth, there is no specific rule of thumb about how to select the most suitable model for a particular breast cancer type and whether that would influence its subsequent application in surgery planning. Typically, these models require tumor biology-dependent parametrization, which hardly generalizes to cope with tumor heterogeneity. In addition, the datasets are limited in size, owing to the restricted or expensive measurement methods. We address the shortcomings that incomplete biological specifications, the variety of tumor types, and the limited size of the data bring to existing mechanistic tumor growth models and introduce a Machine Learning model for the PRediction of INdividual breast Cancer Evolution to Surgical Size (PRINCESS). This is a data-driven model based on neural networks capable of unsupervised learning of cancer growth curves. PRINCESS learns the temporal evolution of the tumor along with the underlying distribution of the measurement space. We demonstrate the superior accuracy of PRINCESS, against four typically used tumor growth models, in learning tumor growth curves from a set of four clinical breast cancer datasets. Our experiments show that, without any modification, PRINCESS can accurately predict tumor sizes while being versatile between breast cancer types.

Virtual Reality (VR) sensorimotor rehabilitation is still in infancy but will soon require avatars, digital alter-egos of patients' physical selves. Such embodied interfaces could stimulate patients' perception in a rich and highly customized environment, where sensorimotor deficits, such as in Chemotherapy-Induced Peripheral Neuropathy, could be corrected. In such scenarios, motion prediction is a key ingredient for realistic immersion. Yet, such a task lives under hard processing latency constraints and the inherent variability of human motion. We propose a neural network meta-learning system exploiting the underlying correlations in body kinematics with potential to provide, within latency guarantees, personalized VR rehabilitation. The unsupervised meta-learner is able to extract underlying statistics of the motion data by exploiting data regularities in order to describe the underlying manifold, or structure, of motion under sensorimotor deficits. We demonstrate, through preliminary experiments the potential of such a learning system for adaptive kinematics estimation in personalized rehabilitation VR avatars.

Remote monitoring health data analysis holds the potential to reduce pregnancy complications, improve patients' quality of life, enhance the efficiency of healthcare delivery and reduce healthcare costs. In this paper, we present a method based on fuzzy inference systems to monitor pregnancies complicated by gestational diabetes mellitus (GDM). The system is simple, fast, flexible and exploits domain expertise in assessing risk levels according to capillary glucose levels from women with GDM. We show that this approach generates an interpretable input, which is valuable in medical applications. To prove the capabilities of the system, we present prediction results from 50 real-world patients and show that the system obtains relevant glycaemic-control data comparable to current monitoring methods that rely on periodic face-to-face physician review. Our systems achieves 95% accuracy. Moreover, we show that the difference in predictions account for a more personalized treatment.

Given the exponential data growth and the recent focus on understanding high-dimensional "in-motion" data, fundamental machine learning tools, such as Principal Component Analysis (PCA), require computation-efficient streaming algorithms that operate near-real-time. Despite the different streaming PCA flavors, there is no algorithm that provably recovers the principal components in the same precision regime as the batch PCA algorithm does, while maintaining low-latency and high-throughput processing. This work, introduces a novel temporal accumulate / retract learning framework for streaming PCA. We consider the accumulate / retract framework implementation of several competitive PCA algorithms with proven theoretical advantages. We benchmark the improved PCA algorithms on real-world streams (i.e. bank transactions fraud detection) and prove their low-latency (millisecond level) and high-throughput (thousands events/second) processing guarantees.

With the evolution of data acquisition methods, our ability to collect real time data has increased. This requires the development of real-time analytics, using the most recent data to generate valuable insights. One example is customer profiling, where we want to identify groups of similar clients who were active recently, and improve the quality of the suggestions. Traditional clustering algorithms perform well on finite datasets, but their execution is often not compatible with real-time requirements, especially for rapid changing trends. In this context, we propose a novel approach for the definition of incremental clustering algorithms to work within real-time constraints, in an online fashion, while preserving accuracy. We show the general applicability of the framework by employing this method to three different clustering algorithms. We compare the experimental results between traditional and online approaches evaluating accuracy and computational cost. The results show that algorithms executed in our framework are comparable to their offline implementation in terms of accuracy and with a high gain in execution time, up to three orders of magnitude on average.

Nowadays, data sources, such as IoT devices, financial markets, and online services, continuously generate large amounts of data. Such data is usually generated at high frequencies and is typically described by non-stationary distributions. Querying these data sources brings new challenges for machine learning algorithms, which now need to be considered from the perspective of an evolving stream and not a static dataset. Under such scenarios, where data flows continuously, the challenge is how to transform the vast amount of data into information and knowledge, and how to adapt to data changes (i.e. drifts) and accumulate experience over time to support online decision-making. In this paper, we introduce STARLORD, a novel incremental computation method and system acting on data streams and capable of achieving low-latency (millisecond level) and high-throughput (thousands events/second/core) when learning from data streams. Moreover, the approach is able to adapt to data drifts and accumulate experience over time, and to use such knowledge to improve future learning and prediction performance, with resource usage guarantees. This is proven by our preliminary experiments where we built-in the framework in an open source stream engine (i.e. Apache Flink).