This Friday the 13th of November at 5pm UTC, Talking Robotics are hosting an online talk with PhD student Natalia Calvo from Uppsala University in Sweden. Talking Robotics is a series of virtual seminars about Robotics and its interaction with other relevant fields, such as Artificial Intelligence, Machine Learning, Design Research, Human-Robot Interaction, among others. The aim is to promote reflections, dialogues, and a place to network. Talking Robotics happens virtually and bi-weekly, i.e., every other week, allocating 30 min for presentation and 30 min for Q&A and networking. Sessions have a roundtable format where everyone is welcome to share ideas.

HRH The Prince of Wales (Front L) Federico Marchetti Front (R) and Back row, members of The Modern ... [ ] Artisan Project team. My recent claim that fashion needs more imagination when it comes to using artificial intelligence has been unexpectedly answered by a project combining eCommerce data and artisanship. Not an obvious pairing, but the brainchild of passionate'dataphile' YOOX NET-A-PORTER GROUP Chairman and CEO, Federico Marchetti, and HRH The Prince of Wales, whose appreciation and support of artisanal craftsmanship (and dedication to safeguarding its future) is decades-long. Marchetti and the YOOX NET-A-PORTER team worked with The Prince's Foundation to create a unique year-long apprenticeship to cultivate the next generation of luxury fashion artisans, informed and guided by customer shopping data and AI analysis of millions of images of historically successful products. To breathe life into artisanship as a viable and attractive career option, underpinned by data that empowers it to deliver the right product, for the right customer on the right sales platform, crucially sustaining the artisans' craft methods and their livelihood.

Domain adaptation solves the learning problem in a target domain by leveraging the knowledge in a relevant source domain. While remarkable advances have been made, almost all existing domain adaptation methods heavily require large amounts of unlabeled target domain data for learning domain invariant representations to achieve good generalizability on the target domain. In fact, in many real-world applications, target domain data may not always be available. In this paper, we study the cases where at the training phase the target domain data is unavailable and only well-labeled source domain data is available, called robust domain adaptation. To tackle this problem, under the assumption that causal relationships between features and the class variable are robust across domains, we propose a novel Causal AutoEncoder (CAE), which integrates deep autoencoder and causal structure learning into a unified model to learn causal representations only using data from a single source domain. Specifically, a deep autoencoder model is adopted to learn low-dimensional representations, and a causal structure learning model is designed to separate the low-dimensional representations into two groups: causal representations and task-irrelevant representations. Using three real-world datasets the extensive experiments have validated the effectiveness of CAE compared to eleven state-of-the-art methods.

Federated learning (FL) is an appealing concept to perform distributed training of Neural Networks (NN) while keeping data private. With the industrialization of the FL framework, we identify several problems hampering its successful deployment, such as presence of non i.i.d data, disjoint classes, signal multi-modality across datasets. In this work, we address these problems by proposing a novel method that not only (1) aggregates generic model parameters (e.g. a common set of task generic NN layers) on server (e.g. in traditional FL), but also (2) keeps a set of parameters (e.g, a set of task specific NN layer) specific to each client. We validate our method on the traditionally used public benchmarks (e.g., Femnist) as well as on our proprietary collected dataset (i.e., traffic classification). Results show the benefit of our method, with significant advantage on extreme cases.

Artificial intelligence is revolutionizing various industries, markets, and services. However, the creative industries and the art world have not yet been able to use the full potential of this technology. However, two Chilean entrepreneurs devised a platform to go further. Using the latest technology, they allow creators, amateur filmmakers, visual artists, even the film and music industry to use artificial intelligence algorithms in their work. This is Runway, a platform that integrates machine learning and artificial intelligence to the world of art and creativity.

To put it simply, the moral circle is the people we care about. Our understanding of it is usually based on William Lecky's History of European Morals from Augustus to Charlemagne. William observes that "at one time the benevolent affections embrace merely the family, soon the circle expanding includes first a class, then a nation, then a coalition of nations, then all humanity, and finally […] the dealings of man with the animal world." In other words, each individual's circle grows as that individual grows older. Just as humanity's moral circle expands from age to age.

In hospitals, it helps doctors decipher X-rays, and in banks, it helps calculate financial risks. Artificial intelligence has been used in commercial relatively recently, but it is no less exciting. But how is make success in such an implementation? From 2012, when one researcher won a competition for image recognition using artificial intelligence, and to this day, when machine learning has become part of the technology industry, all experts understand that MLOps need for the successful use of AI. Cloudera's data engineer, Santiago Giraldo, estimates that only 13% of all projects go through the experimental stage, while others don't. MLOps was created as an analog of DevOps - a modern practice of creating, implementing, and running programs for corporations.

The complete blood count (CBC) performed by automated hematology analyzers is one of the most ordered laboratory tests. It is a first-line tool for assessing a patient's general health status, or diagnosing and monitoring disease progression. When the analysis does not fit an expected setting, technologists manually review a blood smear using a microscope. The International Consensus Group for Hematology Review published in 2005 a set of criteria for reviewing CBCs. Commonly, adjustments are locally needed to account for laboratory resources and populations characteristics. Our objective is to provide a decision support tool to identify which CBC variables are associated with higher risks of abnormal smear and at which cutoff values. We propose a cost-sensitive Lasso-penalized additive logistic regression combined with stability selection. Using simulated and real CBC data, we demonstrate that our tool correctly identify the true cutoff values, provided that there is enough available data in their neighbourhood.

Back in 2019, Hyrum Anderson and I organized the Machine Learning Security Evasion Competition (MLSEC), where participants had to modify malware samples to remain functional and bypass ML-based detection. The competition was successful; the organizers and participants loved it. I get a reply – he's in – let's do this! By the time March rolls around, we are making progress. Our new companies support the competition, and we have ideas for improving it.

We previously proposed a method that allows for nonparallel voice conversion (VC) by using a variant of generative adversarial networks (GANs) called StarGAN. The main features of our method, called StarGAN-VC, are as follows: First, it requires no parallel utterances, transcriptions, or time alignment procedures for speech generator training. Second, it can simultaneously learn mappings across multiple domains using a single generator network and thus fully exploit available training data collected from multiple domains to capture latent features that are common to all the domains. Third, it can generate converted speech signals quickly enough to allow real-time implementations and requires only several minutes of training examples to generate reasonably realistic-sounding speech. In this paper, we describe three formulations of StarGAN, including a newly introduced novel StarGAN variant called "Augmented classifier StarGAN (A-StarGAN)", and compare them in a nonparallel VC task. We also compare them with several baseline methods.