Recently, abstract argumentation-based models of case-based reasoning ($AA{\text -} CBR$ in short) have been proposed, originally inspired by the legal domain, but also applicable as classifiers in different scenarios. However, the formal properties of $AA{\text -} CBR$ as a reasoning system remain largely unexplored. In this paper, we focus on analysing the non-monotonicity properties of a regular version of $AA{\text -} CBR$ (that we call $AA{\text -} CBR_{\succeq}$). Specifically, we prove that $AA{\text -} CBR_{\succeq}$ is not cautiously monotonic, a property frequently considered desirable in the literature. We then define a variation of $AA{\text -} CBR_{\succeq}$ which is cautiously monotonic. Further, we prove that such variation is equivalent to using $AA{\text -} CBR_{\succeq}$ with a restricted casebase consisting of all "surprising" and "sufficient" cases in the original casebase. As a by-product, we prove that this variation of $AA{\text -} CBR_{\succeq}$ is cumulative, rationally monotonic, and empowers a principled treatment of noise in "incoherent" casebases. Finally, we illustrate $AA{\text -} CBR$ and cautious monotonicity questions on a case study on the U.S. Trade Secrets domain, a legal casebase.

Most existing Vision-and-Language (V&L) models rely on pre-trained visual encoders, using a relatively small set of manually-annotated data (as compared to web-crawled data), to perceive the visual world. However, it has been observed that large-scale pretraining usually can result in better generalization performance, e.g., CLIP (Contrastive Language-Image Pre-training), trained on a massive amount of image-caption pairs, has shown a strong zero-shot capability on various vision tasks. To further study the advantage brought by CLIP, we propose to use CLIP as the visual encoder in various V&L models in two typical scenarios: 1) plugging CLIP into task-specific fine-tuning; 2) combining CLIP with V&L pre-training and transferring to downstream tasks. We show that CLIP significantly outperforms widely-used visual encoders trained with in-domain annotated data, such as BottomUp-TopDown. We achieve competitive or better results on diverse V&L tasks, while establishing new state-of-the-art results on Visual Question Answering, Visual Entailment, and V&L Navigation tasks. We release our code at https://github.com/clip-vil/CLIP-ViL.

This article was originally published at The Conversation. The publication contributed the article to Space.com's Astronomy is all about data. The universe is getting bigger and so too is the amount of information we have about it. But some of the biggest challenges of the next generation of astronomy lie in just how we're going to study all the data we're collecting.

Your body is a data goldmine. From the way you look to how you think and feel, firms working in the burgeoning biometrics industry are developing new and alarming ways to track everything we do. And, in many cases, you may not even know you're being tracked. But the biometrics business is on a collision course with Europe's leading data protection experts. Both the European Data Protection Supervisor, which acts as the EU's independent data body, and the European Data Protection Board, which helps countries implement GDPR consistently, have called for a total ban on using AI to automatically recognise people.

There are lots of ML resources around. But so much material how do you pick which ones are good and right for your situation? These are what roadmaps are for. Many people in the ML community have made some you can view. One of the most comprehensive ML roadmaps I have seen. Most beginner to intermediate questions will likely be answered in this mind map.

Parameter reconstruction is a common problem in optical nano metrology. It generally involves a set of measurements, to which one attempts to fit a numerical model of the measurement process. The model evaluation typically involves to solve Maxwell's equations and is thus time consuming. This makes the reconstruction computationally demanding. Several methods exist for fitting the model to the measurements. On the one hand, Bayesian optimization methods for expensive black-box optimization enable an efficient reconstruction by training a machine learning model of the squared sum of deviations. On the other hand, curve fitting algorithms, such as the Levenberg-Marquardt method, take the deviations between all model outputs and corresponding measurement values into account which enables a fast local convergence. In this paper we present a Bayesian Target Vector Optimization scheme which combines these two approaches. We compare the performance of the presented method against a standard Levenberg-Marquardt-like algorithm, a conventional Bayesian optimization scheme, and the L-BFGS-B and Nelder-Mead simplex algorithms. As a stand-in for problems from nano metrology, we employ a non-linear least-square problem from the NIST Standard Reference Database. We find that the presented method generally uses fewer calls of the model function than any of the competing schemes to achieve similar reconstruction performance.

Superconducting electronic circuits have much to offer with regard to neuromorphic hardware. Superconducting quantum interference devices (SQUIDs) can serve as an active element to perform the thresholding operation of a neuron's soma. However, a SQUID has a response function that is periodic in the applied signal. We show theoretically that if one restricts the total input to a SQUID to maintain a monotonically increasing response, a large fraction of synapses must be active to drive a neuron to threshold. We then demonstrate that an active dendritic tree (also based on SQUIDs) can significantly reduce the fraction of synapses that must be active to drive the neuron to threshold. In this context, the inclusion of a dendritic tree provides the dual benefits of enhancing the computational abilities of each neuron and allowing the neuron to spike with sparse input activity.

This paper presents a hybrid morphological neural network for regression tasks called linear dilation-erosion regression ($\ell$-DER). In few words, an $\ell$-DER model is given by a convex combination of the composition of linear and elementary morphological operators. As a result, they yield continuous piecewise linear functions and, thus, are universal approximators. Apart from introducing the $\ell$-DER models, we present three approaches for training these models: one based on stochastic descent gradient and two based on the difference of convex programming problems. Finally, we evaluate the performance of the $\ell$-DER model using 14 regression tasks. Although the approach based on SDG revealed faster than the other two, the $\ell$-DER trained using a disciplined convex-concave programming problem outperformed the others in terms of the least mean absolute error score.

Time-series generated by end-users, edge devices, and different wearables are mostly unlabelled. We propose a method to auto-generate labels of un-labelled time-series, exploiting very few representative labelled time-series. Our method is based on representation learning using Auto Encoded Compact Sequence (AECS) with a choice of best distance measure. It performs self-correction in iterations, by learning latent structure, as well as synthetically boosting representative time-series using Variational-Auto-Encoder (VAE) to improve the quality of labels. We have experimented with UCR and UCI archives, public real-world univariate, multivariate time-series taken from different application domains. Experimental results demonstrate that the proposed method is very close to the performance achieved by fully supervised classification. The proposed method not only produces close to benchmark results but outperforms the benchmark performance in some cases.

But Peter Scott-Morgan has never been afraid of robots. As a scientist and roboticist by trade, he spent decades researching how artificial intelligence (AI) might transform our lives. Then, in 2017, Dr Scott-Morgan was diagnosed with motor neuron disease, the same paralysing condition that killed Stephen Hawking. Months after puzzling over his "wonky foot" falling asleep, he was told he had two years to live. To survive, he would turn to the technology he had spent his career researching.