Recent work has shown that deep reinforcement-learning agents can learn to follow language-like instructions from infrequent environment rewards. However, for many real-world natural language commands that involve a degree of underspecification or ambiguity, such as "tidy the room", it would be challenging or impossible to program an appropriate reward function. To overcome this, we present a method for learning to follow commands from a training set of instructions and corresponding example goal-states, rather than an explicit reward function. Importantly, the example goal-states are not seen at test time. The approach effectively separates the representation of what instructions require from how they can be executed. In a simple grid world, the method enables an agent to learn a range of commands requiring interaction with blocks and understanding of spatial relations and underspecified abstract arrangements. We further show the method allows our agent to adapt to changes in the environment without requiring new training examples.

GuideR: a guided separate-and-conquer rule learning in classification, regression, and survival settings Marek Sikora a,b,, Łukasz Wróbel a,b,, Adam Gudyś a, a Institute of Informatics, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland b Institute of Innovative Technologies, EMAG, Leopolda 31, 40-189 Katowice, PolandAbstract This article presents GuideR, a user-guided rule induction algorithm, which overcomes the largest limitation of the existing methods---the lack of the possibility to introduce user's preferences or domain knowledge to the rule learning process. Automatic selection of attributes and attribute ranges often leads to the situation in which resulting rules do not contain interesting information. We propose an induction algorithm which takes into account user's requirements. Our method uses the sequential covering approach and is suitable for classification, regression, and survival analysis problems. The effectiveness of the algorithm in all these tasks has been verified experimentally, confirming guided rule induction to be a powerful data analysis tool. Introduction Sequential covering rule induction algorithms can be used for both, predictive and descriptive purposes [1, 2, 3, 4]. In spite of the development of increasingly sophisticated versions of those algorithms [5, 6], the main principle remains unchanged and involves two phases: rule growing and rule pruning. In the latter, some of these conditions are removed. In comparison to other machine learning methods, rule sets obtained by sequential covering algorithm, also known as separate-and-conquer strategy (SnC), are characterized by good predictive as well as descriptive capabilities. Taking into consideration only the former, superior results can often be obtained using other methods, e.g. However, data models obtained this way are much less comprehensible than rule sets. In the case of rule learning for descriptive purposes, the algorithms of association rule induction [12, 13, 14] or subgroup discovery [15, 6], are applied. The former leads to a very large number of rules which must then be limited by filtering according to rule interestingness measures [16, 17, 18]. Nevertheless, rule sets obtained by subgroup discovery are characterized by worse predictive abilities than those generated by the standard sequential covering approach. Therefore, if creating a prediction system with comprehensible data model is the main objective, the application of sequential covering rule induction algorithms provides the most sensible solution.

Summary: A new artificial intelligence technique could speed up complex physics simulations and help create multilayered nanoparticles, researchers say. A new technique developed by MIT physicists could someday provide a way to custom-design multilayered nanoparticles with desired properties, potentially for use in displays, cloaking systems, or biomedical devices. It may also help physicists tackle a variety of thorny research problems, in ways that could in some cases be orders of magnitude faster than existing methods. The innovation uses computational neural networks, a form of artificial intelligence, to "learn" how a nanoparticle's structure affects its behavior, in this case the way it scatters different colors of light, based on thousands of training examples. Then, having learned the relationship, the program can essentially be run backward to design a particle with a desired set of light-scattering properties -- a process called inverse design.

Digital texts have become an increasingly important source of data for social studies. However, textual data from open platforms are vulnerable to manipulation (e.g., censorship and information inflation), often leading to bias in subsequent empirical analysis. This paper investigates the problem of data distortion in text classification when controlling type I error (a relevant textual message is classified as irrelevant) is the priority. The default classical classification paradigm that minimizes the overall classification error can yield an undesirably large type I error, and data distortion exacerbates this situation. As a solution, we propose the Neyman-Pearson (NP) classification paradigm which minimizes type II error under a user-specified type I error constraint. Theoretically, we show that while the classical oracle (i.e., optimal classifier) cannot be recovered under unknown data distortion even if one has the entire post-distortion population, the NP oracle is unaffected by data distortion and can be recovered under the same condition. Empirically, we illustrate the advantage of NP classification methods in a case study that classifies posts about strikes and corruption published on a leading Chinese blogging platform.

The innovation uses computational neural networks, a form of artificial intelligence, to "learn" how a nanoparticle's structure affects its behavior, in this case the way it scatters different colors of light, based on thousands of training examples. Then, having learned the relationship, the program can essentially be run backward to design a particle with a desired set of light-scattering properties -- a process called inverse design. The findings are being reported in the journal Science Advances, in a paper by MIT senior John Peurifoy, research affiliate Yichen Shen, graduate student Li Jing, professor of physics Marin Soljacic, and five others. While the approach could ultimately lead to practical applications, Soljacic says, the work is primarily of scientific interest as a way of predicting the physical properties of a variety of nanoengineered materials without requiring the computationally intensive simulation processes that are typically used to tackle such problems. Soljacic says that the goal was to look at neural networks, a field that has seen a lot of progress and generated excitement in recent years, to see "whether we can use some of those techniques in order to help us in our physics research. So basically, are computers'intelligent' enough so that they can do some more intelligent tasks in helping us understand and work with some physical systems?"

Which authors of this paper are endorsers? Disable MathJax (What is MathJax?)

Constraint-based learning reduces the burden of collecting labels by having users specify general properties of structured outputs, such as constraints imposed by physical laws. We propose a novel framework for simultaneously learning these constraints and using them for supervision, bypassing the difficulty of using domain expertise to manually specify constraints. Learning requires a black-box simulator of structured outputs, which generates valid labels, but need not model their corresponding inputs or the input-label relationship. At training time, we constrain the model to produce outputs that cannot be distinguished from simulated labels by adversarial training. Providing our framework with a small number of labeled inputs gives rise to a new semi-supervised structured prediction model; we evaluate this model on multiple tasks --- tracking, pose estimation and time series prediction --- and find that it achieves high accuracy with only a small number of labeled inputs. In some cases, no labels are required at all.

A pair of academics have reproduced part of a moth's brain as an artificial neural network – and taught it to recognize numbers to a fairly high accuracy with just a few training examples. The software, dubbed MothNet, can apparently discern handwritten digits with 75 per cent to 85 per cent accuracy, given 15 to 20 training samples of each number. That's not bad considering it takes thousands of training examples for more traditional neural networks to achieve 99 per cent accuracy. Its masterminds, Charles B. Delahunt and J. Nathan Kutz, both at the University of Washington in the US, built MothNet by modeling the olfactory network – the part of the brain that processes smells – found in the Carolina sphinx moth, also known as the tobacco hawk moth (Manduca sexta). That section of a bug's grey matter is relatively straightforward, we're told, making it ideal for experimentation.

Which authors of this paper are endorsers? Disable MathJax (What is MathJax?)

Real-world machine learning applications often have complex test metrics, and may have training and test data that follow different distributions. We propose addressing these issues by using a weighted loss function with a standard convex loss, but with weights on the training examples that are learned to optimize the test metric of interest on the validation set. These metric-optimized example weights can be learned for any test metric, including black box losses and customized metrics for specific applications. We illustrate the performance of our proposal with public benchmark datasets and real-world applications with domain shift and custom loss functions that balance multiple objectives, impose fairness policies, and are non-convex and non-decomposable.