After 12 years of activity, NCCR Robotics officially ended on 30 November 2022. We can proudly say that NCCR Robotics has had a truly transformational effect on the national robotics research landscape, creating novel synergies, strengthening key areas, and adding a unique signature that made Switzerland prominent and attractive at the international level. In its 12 years of activity, NCCR Robotics has had a transformational effect on the national robotics research landscape. It is not the end of the story though: our partner institutions – EPFL, ETH Zurich, the University of Zurich, the University of Bern, the University of Basel, Università della Svizzera Italiana, EMPA – will continue to collaborate through the Innovation Booster Robotics, a new national program aimed at developing technology transfer activities and maintaining the network. The research programme of NCCR Robotics has been articulated around three Grand Challenges for future intelligent robots that can improve the quality of life: Wearable Robotics, Rescue Robotics, and Educational Robotics.

The Introduction to Robotics Specialization introduces you to the concepts of robot flight and movement, how robots perceive their environment, and how they adjust their movements to avoid obstacles, navigate difficult terrains and accomplish complex tasks such as construction and disaster recovery. You will be exposed to real world examples of how robots have been applied in disaster situations, how they have made advances in human health care and what their future capabilities will be. The courses build towards a capstone in which you will learn how to program a robot to perform a variety of movements such as flying and grasping objects.

Fancy joining a global organisation that is revolutionizing the digital landscape? Today, as clients across industries are moving from digitally extending their businesses to placing digital at the core, Publicis Sapient has an unprecedented opportunity to help them succeed. As a Senior Engineer with our Data Engineering group, you will be responsible for the design and implementation of high-end software products/services that enable the enterprise-scale digital transformation of many of the biggest companies in the world. Your role will be focused on delivering solutions that leverage large scale data ingestion, processing, storage/querying, in-stream & batch analytics. You will also be involved in building technology prototypes for validation and assess technical designs for functional & non-functional completeness.

There is no doubt that Machine Learning is a tough subject, and in-depth knowledge, in particular, requires a lot of Mathematics and complex terminology and is very tough to master. How do you learn it better if the subject matter is that tough? Choose a course that can explain this complex topic in simple words. We are actually blessed that we have many excellent instructors like Andrew Ng, Jeremey Howard, and Kirill Eremenko on Udemy, who are not just experts in deep learning but also excellent instructors and teachers. I firmly believe that every programmer should learn about Cloud Computing and Artificial Intelligence, as these two will drive the world in the coming years.

This three-module course introduces machine learning and data science for everyone with a foundational understanding of machine learning models. You'll learn about the history of machine learning, applications of machine learning, the machine learning model lifecycle, and tools for machine learning. You'll also learn about supervised versus unsupervised learning, classification, regression, evaluating machine learning models, and more. Our labs give you hands-on experience with these machine learning and data science concepts. You will develop concrete machine learning skills as well as create a final project demonstrating your proficiency.

Machine Learning for Source Code (ML4Code) is an active research field in which extensive experimentation is needed to discover how to best use source code's richly structured information. With this in mind, we introduce JEMMA, an Extensible Java Dataset for ML4Code Applications, which is a large-scale, diverse, and high-quality dataset targeted at ML4Code. Our goal with JEMMA is to lower the barrier to entry in ML4Code by providing the building blocks to experiment with source code models and tasks. JEMMA comes with a considerable amount of pre-processed information such as metadata, representations (e.g., code tokens, ASTs, graphs), and several properties (e.g., metrics, static analysis results) for 50,000 Java projects from the 50KC dataset, with over 1.2 million classes and over 8 million methods. JEMMA is also extensible allowing users to add new properties and representations to the dataset, and evaluate tasks on them. Thus, JEMMA becomes a workbench that researchers can use to experiment with novel representations and tasks operating on source code. To demonstrate the utility of the dataset, we also report results from two empirical studies on our data, ultimately showing that significant work lies ahead in the design of context-aware source code models that can reason over a broader network of source code entities in a software project, the very task that JEMMA is designed to help with.

A hallmark of human intelligence is the ability to learn new concepts purely from language. Several recent approaches have explored training machine learning models via natural language supervision. However, these approaches fall short in leveraging linguistic quantifiers (such as 'always' or 'rarely') and mimicking humans in compositionally learning complex tasks. Here, we present LaSQuE, a method that can learn zero-shot classifiers from language explanations by using three new strategies - (1) modeling the semantics of linguistic quantifiers in explanations (including exploiting ordinal strength relationships, such as 'always' > 'likely'), (2) aggregating information from multiple explanations using an attention-based mechanism, and (3) model training via curriculum learning. With these strategies, LaSQuE outperforms prior work, showing an absolute gain of up to 7% in generalizing to unseen real-world classification tasks.

In reinforcement learning applications like robotics, agents usually need to deal with various input/output features when specified with different state/action spaces by their developers or physical restrictions. This indicates unnecessary re-training from scratch and considerable sample inefficiency, especially when agents follow similar solution steps to achieve tasks. In this paper, we aim to transfer similar high-level goal-transition knowledge to alleviate the challenge. Specifically, we propose PILoT, i.e., Planning Immediate Landmarks of Targets. PILoT utilizes the universal decoupled policy optimization to learn a goal-conditioned state planner; then, distills a goal-planner to plan immediate landmarks in a model-free style that can be shared among different agents. In our experiments, we show the power of PILoT on various transferring challenges, including few-shot transferring across action spaces and dynamics, from low-dimensional vector states to image inputs, from simple robot to complicated morphology; and we also illustrate a zero-shot transfer solution from a simple 2D navigation task to the harder Ant-Maze task.

Variational inference (VI) is a specific type of approximate Bayesian inference that approximates an intractable posterior distribution with a tractable one. VI casts the inference problem as an optimization problem, more specifically, the goal is to maximize a lower bound of the logarithm of the marginal likelihood with respect to the parameters of the approximate posterior. Reinforcement learning (RL) on the other hand deals with autonomous agents and how to make them act optimally such as to maximize some notion of expected future cumulative reward. In the non-sequential setting where agents' actions do not have an impact on future states of the environment, RL is covered by contextual bandits and Bayesian optimization. In a proper sequential scenario, however, where agents' actions affect future states, instantaneous rewards need to be carefully traded off against potential long-term rewards. This manuscript shows how the apparently different subjects of VI and RL are linked in two fundamental ways. First, the optimization objective of RL to maximize future cumulative rewards can be recovered via a VI objective under a soft policy constraint in both the non-sequential and the sequential setting. This policy constraint is not just merely artificial but has proven as a useful regularizer in many RL tasks yielding significant improvements in agent performance. And second, in model-based RL where agents aim to learn about the environment they are operating in, the model-learning part can be naturally phrased as an inference problem over the process that governs environment dynamics. We are going to distinguish between two scenarios for the latter: VI when environment states are fully observable by the agent and VI when they are only partially observable through an observation distribution.

Data Visualization is new Analytics and, Augmented Analytics is new Data Visualization! In this course you will work on machine learning models for predictive analytics and advanced data flow features through hands on training with Oracle Analytics. This course is designed to provide you with many hands-on activities to learn building modern data visualization projects. This is new business intelligence! Are you a business analyst curious about what Oracle Analytics can do?