Search

We propose a filtering feature selection framework that considers subsets of features as paths in a graph, where a node is a feature and an edge indicates pairwise (customizable) relations among features, dealing with relevance and redundancy principles. By two different interpretations (exploiting properties of power series of matrices and relying on Markov chains fundamentals) we can evaluate the values of paths (i.e., feature subsets) of arbitrary lengths, eventually go to infinite, from which we dub our framework Infinite Feature Selection (Inf-FS). Going to infinite allows to constrain the computational complexity of the selection process, and to rank the features in an elegant way, that is, considering the value of any path (subset) containing a particular feature. We also propose a simple unsupervised strategy to cut the ranking, so providing the subset of features to keep. In the experiments, we analyze diverse settings with heterogeneous features, for a total of 11 benchmarks, comparing against 18 widely-known comparative approaches. The results show that Inf-FS behaves better in almost any situation, that is, when the number of features to keep are fixed a priori, or when the decision of the subset cardinality is part of the process., Comment: TPAMI PREPRINT 2020

Insert here your abstract text.

One of the main and most effective measures to contain the recent viral outbreak is the maintenance of the so-called Social Distancing (SD). To comply with this constraint, workplaces, public institutions, transports and schools will likely adopt restrictions over the minimum inter-personal distance between people. Given this actual scenario, it is crucial to massively measure the compliance to such physical constraint in our life, in order to figure out the reasons of the possible breaks of such distance limitations, and understand if this implies a possible threat given the scene context. All of this, complying with privacy policies and making the measurement acceptable. To this end, we introduce the Visual Social Distancing (VSD) problem, defined as the automatic estimation of the inter-personal distance from an image, and the characterization of the related people aggregations. VSD is pivotal for a non-invasive analysis to whether people comply with the SD restriction, and to provide statistics about the level of safety of specific areas whenever this constraint is violated. We then discuss how VSD relates with previous literature in Social Signal Processing and indicate which existing Computer Vision methods can be used to manage such problem. We conclude with future challenges related to the effectiveness of VSD systems, ethical implications and future application scenarios., 9 pages, 5 figures. All the authors equally contributed to this manuscript and they are listed by alphabetical order. Under submission

In this paper we discuss a transition model from commonly adopted models of data gathering, transfer and management for environmental monitoring towards more sophisticated ones based on Artificial Intelligence and IoT. The transition model is based on the paradigm of multiple agent systems. The adoption of this transition model is motivated by the need to improve effectiveness, efficiency and interoperability of environmental monitoring by simultaneously guaranteeing its sustainability in economic terms.

The use of meta-rules, i.e., rules whose content includes other rules, has been advocated to model policies and the notion of power in legal reasoning, where an agent has the power to create new norms affecting other agents. The use of Defeasible Logic (DL) to model meta-rules in the application area we just alluded to has been investigated, but not from a computational viewpoint. Our aim is to fill this gap by introducing a variant of DL, Defeasible Meta-Logic, to represent defeasible meta-theories, by proposing efficient algorithms to compute the (meta-)extensions of such theories, and by proving their computational complexity.

This paper presents DOHMO, an embedded computer vision system where multiple sensors, including intelligent cameras, are connected to actuators that regulate illumination and doors. The system aims at assisting elderly and impaired people in co-housing scenarios, in accordance with privacy design principles. The paper provides details of two core elements of the system: The first one is the BOX-IO controller, a fully scalable and customizable hardware and software IoT ecosystem that can collect, control, and monitor data, operational flows and business scenarios, whether indoor or outdoor. The second one is the embedded 3DEverywhere intelligent camera, a device composed of an embedded system that receives input data provided by a 3D/2D camera, analyzes it, and returns the metadata of this analysis. We illustrate how they can be connected and how simple decision mechanisms can be implemented in such a framework. In particular, illumination can be triggered on and off by the detected presence of people, overcoming the limitations of typical sensors, while doors can be opened or closed based on person trajectories in an intelligent manner. To substantiate the proposed system, numerous experiments are performed in a lab and a co-housina scenario.

In this work, we describe the main processes by which an agent operating in a normative/ethical system may formulate a new normative/ethical rule. This is an issue that emerges as fundamental in hybrid societies, where humans and autonomous agents shall live together. In these environments, it may occur a situation that is totally unknown to the agent, who may be unable to apply existing rule and need to create new ones to deal with the novelties encountered. We employ, as a means to devise the revision operators, the defeasible deontic logic (DDL), a non-monotonic logical system that accommodates rules of ethical origin, and has been widely employed for normative and moral reasoning. New rules may be acquired by an external source, possibly an institution, or a human interacting with an intelligent agent. The agent itself may learn it in a number of manners. We introduce two revision operators for norms that are able to manage the effects of all the above-mentioned processes.

In this paper, we propose the use of a new modality characterized by a richer information content, namely acoustic images, for the sake of audio-visual scene understanding. Each pixel in such images is characterized by a spectral signature, associated to a specific direction in space and obtained by processing the audio signals coming from an array of microphones. By coupling such array with a video camera, we obtain spatio-temporal alignment of acoustic images and video frames. This constitutes a powerful source of self-supervision, which can be exploited in the learning pipeline we are proposing, without resorting to expensive data annotations. However, since 2D planar arrays are cumbersome and not as widespread as ordinary microphones, we propose that the richer information content of acoustic images can be distilled, through a self-supervised learning scheme, into more powerful audio and visual feature representations. The learnt feature representations can then be employed for downstream tasks such as classification and cross-modal retrieval, without the need of a microphone array. To prove that, we introduce a novel multimodal dataset consisting in RGB videos, raw audio signals and acoustic images, aligned in space and synchronized in time. Experimental results demonstrate the validity of our hypothesis and the effectiveness of the proposed pipeline, also when tested for tasks and datasets different from those used for training.

Data anonymization is the process of making information contained in a group of data such that it is not possible to identify unique references to single elements in the group after the process. This action, when conducted onto datasets used to make statistical inference is bound to have ananlogous behaviours on certain indices before and after the process itself. In this paper we study the pipeline of anonymization process for datasets, when this pipeline is managed on cloud technology, where cryptography is not applicable at all, for datasets being available in an open setting. We examine the open problems, and devise a method to address these problems in a logical framework.

Defeasible deontic logic has shown to be expressive enough to represent a normative system, and therefore compliance to such a system can be automatically checked by means of classical model checking techniques of logical systems. However, normative systems are not static, as they can be actively changed by the legislator over time, directly, by changing one norm. Moreover norms can change passively, either by effect of the change of another piece of the normative system, or by means of the change of meaning that affects terms employed in the norm.

Agents use channels to communicate publicly. An agent announcing a statement on a communication channel poses non trivial questions about the nature of such a statement as well as about the attitude of the agent herself. Does the agent know whether the statement is true? Is this agent announcing that statement or its contrary in any other channel?

Some wearable solutions exploiting on-body acceleration sensors have been proposed to recognize Freezing of Gait (FoG) in people affected by Parkinson Disease (PD). Once a FoG event is detected, these systems generate a sequence of rhythmic stimuli to allow the patient restarting the gait. While these solutions are effective in detecting FoG events, they are unable to predict FoG to prevent its occurrence. This paper fills in the gap by presenting a machine learning-based approach that classifies accelerometer data from PD patients, recognizing a pre-FOG phase to further anticipate FoG occurrence in advance. Gait was monitored by three tri-axial accelerometer sensors worn on the back, hip and ankle. Gait features were then extracted from the accelerometer's raw data through data windowing and non-linear dimensionality reduction. A k-nearest neighbor algorithm (k-NN) was used to classify gait in three classes of events: pre-FoG, no-FoG and FoG. The accuracy of the proposed solution was compared to state-of-the-art approaches. Our study showed that: (i) we achieved performances overcoming the state-of-the-art approaches in terms of FoG detection, (ii) we were able, for the very first time in the literature, to predict FoG by identifying the pre-FoG events with an average sensitivity and specificity of, respectively, 94.1% and 97.1%, and (iii) our algorithm can be executed on resource-constrained devices. Future applications include the implementation on a mobile device, and the administration of rhythmic stimuli by a wearable device to help the patient overcome the FoG.

We propose a low-cost, open-source platform to encourage large-scale study and research on human-robot social interaction. The paramount of social interaction lies on face-to-face dynamics, so we focus on realizing an anthropomorphic robotic head, Berrick, with multimodal sensing and acting capabilities. Taking from the InMoov robot [1], we isolate its head and redesign its electronics, exploiting the implementation of a new board able to combine an autonomous vision system with the already present motorized facial platform, with Wi-Fi and Bluetooth connectivity for multiagent communication. At the present moment, Berrick is capable of running fundamental yet sophisticated tasks of face detection and gazing, which are crucial to trigger and drive situated social exchanges. In particular, we present here a novel social cue embedded into Berrick, dubbed light-based gazing, functional to communicate the internal state of the robot during a dyadic interaction, thus facilitating situated social exchanges. With a 250€ cost, a 3 days time-to-build (with fully 3D printable parts) and a publicly available documentation, Berrick can become a reference for approaching human robot social interaction at a large-scale at the universities as well as at lower education degrees.

A semantic community is a set of individuals who are interested in the contents that refer to a specific domain, around which they aggregate to perform social activities such as sharing, commenting, possibly editing, those contents. Every producer of contents that is interested in interacting on social networks and in general on digital news platforms is interested in actively collaborating with the semantic communities that exist about the topics she produces. In general, moreover, a content producer is interested in fostering communities, mainly because this will generate a higher interest on her contents. In this paper, we illustrate an architecture of a system able to support and manage semantic communities in the framework of a project for digital news delivering. The architecture is illustrated and its basic concept is presented.

In shape analysis and matching, it is often important to encode information about the relation between a given point and other points on a shape, namely, its context . To this aim, we propose a theoretically sound and efficient approach for the simulation of a discrete time evolution process that runs through all possible paths between pairs of points on a surface represented as a triangle mesh in the discrete setting. We demonstrate how this construction can be used to efficiently construct a multiscale point descriptor, called the Discrete Time Evolution Process Descriptor , which robustly encodes the structure of neighborhoods of a point across multiple scales. Our work is similar in spirit to the methods based on diffusion geometry, and derived signatures such as the HKS or the WKS, but provides information that is complementary to these descriptors and can be computed without solving an eigenvalue problem. We demonstrate through extensive experimental evaluation that our descriptor can be used to obtain accurate results in shape matching in different scenarios. Our approach outperforms similar methods and is especially robust in the presence of large nonisometric deformations, including missing parts.

In this paper we introduce a simplified approach to sentiment analysis: a lexicon-driven method based upon only adjectives and adverbs. This method is compared in cross-validation with other known techniques and then compared directly to the gold standard, a sample of human subjects asked to deliver the same class of judgments computed by the method. We prove that the method is similar in accuracy and precision with the other methods. We finally argue that the approach we employ is more valid than others for it is scalable, and exportable to languages other than English.

Normative systems accommodate temporary norms of several types, which can also be modified in different, and codified ways. In this paper we address the problem of modifying temporary norms that are represented by means of the combination of two known formalisms in the current literature. The framework evolves from a known one, which provides a system of norms at two distinct layers, and represents changes at the two layers as means to provide room for the codified change types. This results in four novel operators that anticipate and extend norms in two different combined ways, by preserving or not the effects of the norms in the period of time generated by the temporal modifications. We study these new oper- ators and show how they relate to the operators of annulment and abrogation analysed elsewhere.

In this tutorial we describe the approaches to non monotonic reasoning as a means for inference on the web. In particular we are focusing on the ways in which reasoning technologies have adapted to five different issues of the modern era world wide web: (a) epistemic aspects, bound by the new models of the social web, (b) changes over time, (c) language variants, including different languages of deployment of a web site, (d) agent-based knowledge deployment, due to social networks and blogs, (e) dialogue aspects, introduced again in blogs and social networks. The presentation covers these aspects by a technical viewpoint, including the introduction of specific knowledge-driven methods. The technical issues will be provided within a general logical framework known as defeasible logic.

In this paper, we illustrate a combinatorial approach to the problem of defining an ontology used for detecting multiple authorship and roles in weblogs and social network literature production. The investigation aims at providing a model of workflows and therefore traits to be recognised in these documents.

We present a new paradigm for human-robot interaction based on social signal processing, and in particular on the Brunswick model. Originally, the Brunswick model copes with face-to-face dyadic interaction, assuming that the interactants are communicating through a continuous exchange of non verbal social signals, in addition to the spoken messages. Social signals have to be interpreted, thanks to a proper recognition phase that considers visual and audio information. The Brunswick model allows to quantitatively evaluate the quality of the interaction using statistical tools which measure how effective is the recognition phase. In this paper we cast this theory when one of the interactants is a robot; in this case, the recognition phase performed by the robot and the human have to be revised w.r.t. the original model. The model is applied to Berrick, a recent open-source low-cost robotic head platform, where the gazing is the social signal to be considered.

The detection of groups of individuals is attracting the attention of many researchers in diverse fields, from automated surveillance to human-computer interaction, with a growing number of approaches published every year. Unexpectedly, the evaluation metrics for this problem are not consolidated, with some measures inherited from the people detection field, other from clustering, other designed specifically for a particular approach, thus lacking in generalization and making the comparisons between different approaches hard to be carried out. Moreover, most of the existent metrics are scarcely expressive, addressing groups as they are atomic entities, ignoring that they may have different cardinalities, and that group detection approaches may fail in capturing the exact number of individuals that compose it. This paper fills this gap presenting the GROup DEtection (GRODE) metrics, which formally define precision and recall on the groups, including the group cardinality as a variable. This gives the possibility to investigate aspects never considered so far, such as the tendency of a method of over- or under-segmenting, or of better dealing with specific group cardinalities. The GRODE metrics have been evaluated first on controlled scenarios, where the differences with alternative metrics are evident. Then, the metrics have been applied to eight approaches of group detection, on eight public datasets, providing a fresh-new panorama of the state-of-the-art, discovering interesting strengths and pitfalls of the recent approaches.

Linear Logic and Defeasible Logic have been adopted to formalise different features of knowledge representation: consumption of resources, and non monotonic reasoning in particular to represent exceptions. Recently, a framework to combine sub-structural features, corresponding to the consumption of resources, with defeasibility aspects to handle potentially conflicting information, has been discussed in literature, by some of the authors. Two applications emerged that are very relevant: energy management and business process management. We illustrate a set of guide lines to determine how to apply linear defeasible logic to those contexts., Comment: In Proceedings DICE-FOPARA 2019, arXiv:1908.04478. arXiv admin note: substantial text overlap with arXiv:1809.03656

In this paper we propose an image segmentation method specifically designed to detect crystalline grains in microscopic images. We build on the watershed segmentation approach; we propose a preprocessing pipeline to generate a topographic map exploiting the physical nature of the incoming data (i.e. Atomic Force Microscopy) to emphasize grain boundaries and generate seeds for basins. Experimental results show the effectiveness of the proposed method against grain segmentation implementations available in commercial software on a new labelled dataset with an average improvement of over 20% in precision and recall over the standard implementation of watershed segmentation.

Text normalisation is an important task in the context of Natural Language Processing. By normalisation, free text is mapped into dictionaries, i.e. indexed collections of locutions recognised as typical of a particular jaergon. In general, technical dictionaries are difficult to build and validate. They are typically constructed by hand on the basis of everyday human work and they are agreement-based. This is indubitably time consuming and the approach requires a strong human supervision and does not provide a general methodology. In this paper, we perform the first steps towards the to automatic building of a dictionary for Italian journalistic lexicon, called NewsDict, based on sub dictionaries able to characterise main topics occurring in newspaper articles. We exploit a dataset of annotated documents from some Italian newspapers and a statistical techniques based on the Mutual Information Principle. Documents contains information such as the release date and the topic of the article and has been directly annotated by the author. To check the accuracy of the dictionary we built, we develop an initial test. We normalise a control set of journal article into NewsDict. Crossing results presented in this paper against the human annotation, we provide a fist measure of performances of the described methodology.

There are several contexts of non-monotonic reasoning where a priority between rules is established with the purpose of solving conflicts. We investigate how to modify such a priority (preference) relation in a non-monotonic logic in order to change the conclusions of the theory itself. We shall argue that the approach we adopt has a natural counterpart in legal reasoning and argumentation, where users cannot typically change the facts or the rules, but can propose their preferences about the relative strength of the rules. The main result of the present work is the proof that the problem of revising a non-monotonic theory by changing only the superiority order between conflicting rules is, in general, computationally hard. After such an analysis, we identify three contraction/revision/update operations and study them against the AGM postulates for belief revision, to discover that only a (small) part of these postulates are satisfied in the specific non-monotonic setting.

The social choice theory has focused in the past on the problem of devising methods to determine how individual preferences are transformed into collective ones. In some investigations, scholars provided methods for expressing the social choice function, that, given a set of individual preferences, computes the resulting collective choice. Other studies focused on determining under which conditions the social choice function is efficiently computable.

Object pose estimation is important for systems and robots to interact with the environment where the main challenge of this task is the complexity of the scene caused by occlusions and clutters. A key challenge is performing pose estimation leveraging on both RGB and depth information: prior works either extract information from the RGB image and depth separately or use costly post-processing steps, limiting their performances in highly cluttered scenes and real-time applications. Traditionally, the pose estimation problem is tackled by matching feature points between 3D models and images. However, these methods require rich textured models. In recent years, the raising of deep learning has offered an increasing number of methods based on neural networks, such as DSAC++, PoseCNN, DenseFusion and SingleShotPose. In this work, we present a comparison between two recent algorithms, DSAC++ and DenseFusion, focusing on computational cost, performance and applicability in the industry.

We present SIMCO, the first agnostic multi-class object counting approach. SIMCO starts by detecting foreground objects through a novel Mask RCNN-based architecture trained beforehand (just once) on a brand-new synthetic 2D shape dataset, InShape; the idea is to highlight every object resembling a primitive 2D shape (circle, square, rectangle, etc.). Each object detected is described by a low-dimensional embedding, obtained from a novel similarity-based head branch; this latter implements a triplet loss, encouraging similar objects (same 2D shape + color and scale) to map close. Subsequently, SIMCO uses this embedding for clustering, so that different types of objects can emerge and be counted, making SIMCO the very first multi-class unsupervised counter. Experiments show that SIMCO provides state-of-the-art scores on counting benchmarks and that it can also help in many challenging image understanding tasks., Accepted at ICPR 2020

Capturing the essence of a textile image in a robust way is important to retrieve it in a large repository, especially if it has been acquired in the wild (by taking a photo of the textile of interest). In this paper we show that a texel-based representation fits well with this task. In particular, we refer to Texel-Att, a recent texel-based descriptor which has shown to capture fine grained variations of a texture, for retrieval purposes. After a brief explanation of Texel-Att, we will show in our experiments that this descriptor is robust to distortions resulting from acquisitions in the wild by setting up an experiment in which textures from the ElBa (an Element-Based texture dataset) are artificially distorted and then used to retrieve the original image. We compare our approach with existing descriptors using a simple ranking framework based on distance functions. Results show that even under extreme conditions (such as a down-sampling with a factor of 10), we perform better than alternative approaches., Comment: ICIAP - International Conference on Image Analysis and Processing

Since the birth of digital social networks, management research focused upon the opportunities of social media marketing. A marketing campaign has the best success when it reaches the largest number of potential customers. It is, however, difficult to forecast in a precise way the number of contacts that you can reach with such an initiative.

Linear Logic and Defeasible Logic have been adopted to formalise different features of knowledge representation: consumption of resources and reasoning with exceptions. We propose a framework to combine sub-structural features, corresponding to the consumption of resources, with defeasibility aspects to handle potentially conflicting information, and we discuss the design choices.

Weather forecasting is a logical process that consists in evaluating the predictions provided by a set of stochastic models, compare these and take a conclusion about the weather in a given area and a given interval of time. Meteorological forecasting provides reliable predictions about the weather within a given interval of time. The automation of the forecasting process would be helpful in a number of contexts. For instance, when forecasting about underpopulated or small geographic areas is out of the human forecasters’ tasks but is central, e.g., for tourism. In this paper, we start to deal with these challenging tasks by developing a defeasible reasoner for meteorological forecasting, which we evaluate against a real-world example with applications to tourism and holiday planning.

Lighting design and modelling or industrial applications like luminaire planning and commissioning rely heavily on time consuming manual measurements or on physically coherent computational simulations. Regarding the latter,standard approaches are based on CAD modeling simulations and offline rendering, with long processing times and therefore inflexible workflows. Thus, in this paper we pro-pose a computer vision based system to measure lighting with just a single RGBD camera. The proposed method uses both depth data and images from the sensor to provide a dense measure of light intensity in the field of view of the camera. We evaluate our system on novel ground truth data and compare it to state-of-the-art commercial light-planning software. Our system provides improved performance, while being completely automated, given that the CAD model is extracted from the depth and the albedo estimated with the support of RGB images. To the best of our knowledge, this is the first automatic framework for the estimation of lighting in general indoor scenarios from RGBDinput., 10 pages, 9 figures, this manuscript is accepted in WACV 2019

This work deals with the problem of automatic detection of device types given only the power consumption curve, which can be obtained by means of a cheap measurer applied to the device itself. We defined a novel method to detect these types and we describe it in details, providing ground truth evidence coming from the application of the method to real world data. We tested the method against two different set of data coming from two separate and different environments, the first located in Italy and the second in Germany, and we provide experimental results to support the method.

Diagnosis is widely used in many different disciplines to identify the nature and cause of a certain phenomenon. We present \(t\mathsf {L}\), a new logical framework able to formalise diagnostic reasoning, i.e., an hybrid learning technique based both on deduction and experiments. In this paper we introduce tL, a Labeled Modal Logic, garnishing with temporal and statistical information and a basic propositional language.

Agents negotiate the meaning of terms in numerous real-life situations. When they behave so, they can be used as the basis for providing an emulation paradigm for software agents, habilitating therefore socio-technical systems to perform meaning negotiation. In this paper we focus upon two methods for meaning negotiation in defeasible logic and provide room for an analysis of how the proposed approaches perform the aforementioned process. Finally we also provide a computational analysis of the process automation problem.

The frequency and harmfulness of cyber-attacks are increas- ing every day, and with them also the amount of data that the cyber- forensics analysts need to collect and analyze. In this paper, we pro- pose a formal analysis process that allows an analyst to filter the enor- mous amount of evidence collected and either identify crucial informa- tion about the attack (e.g., when it occurred, its culprit, its target) or, at the very least, perform a pre-analysis to reduce the complexity of the problem in order to then draw conclusions more swiftly and efficiently. We introduce the Evidence Logic EL for representing simple and derived pieces of evidence from different sources. We propose a procedure, based on monotonic reasoning, that rewrites the pieces of evidence with the use of tableau rules, based on relations of trust between sources and the reasoning behind the derived evidence, and yields a consistent set of pieces of evidence. As proof of concept, we apply our analysis process to a concrete cyber-forensics case study.

In this paper we show the importance of the head pose estimation in the task of trajectory forecasting. This cue, when produced by an oracle and injected in a novel socially-based energy minimization approach, allows to get state-of-the-art performances on four different forecasting benchmarks, without relying on additional information such as expected destination and desired speed, which are supposed to be know beforehand for most of the current forecasting techniques. Our approach uses the head pose estimation for two aims: 1) to define a view frustum of attention, highlighting the people a given subject is more interested about, in order to avoid collisions; 2) to give a shorttime estimation of what would be the desired destination point. Moreover, we show that when the head pose estimation is given by a real detector, though the performance decreases, it still remains at the level of the top score forecasting systems.

The head pose is an important cue for computer vision. Traditionally considered in human computer interaction applications, it becomes very hard to model in surveillance scenarios, due to the tiny head size. Additionally, no public dataset contains continuous head pose annotations in open scenery, making the challenge even harder to face. Here we present a framework based on Faster RCNN, which introduces a branch in the network architecture related to the head pose estimation. The key idea is to leverage the presence of the people body to better infer the head pose, through a joint optimization process. Additionally, we enrich the Town Center dataset with head pose labels, promoting further study on this topic. Results on this novel benchmark and ablation studies on other task-specific datasets promote our idea and confirm the importance of the body cues to contextualize the head pose estimation.

Linear Logic and Defeasible Logic have been adopted to formalise different features relevant to agents: consumption of resources, and reasoning with exceptions. We propose a framework to combine sub-structural features, corresponding to the consumption of resources, with defeasibility aspects, and we discuss the design choices for the framework.

Lighting design in indoor environments is of primary importance for at least two reasons: 1) people should perceive an adequate light; 2) an effective lighting design means consistent energy saving. We present the Invisible Light Switch (ILS) to address both aspects. ILS dynamically adjusts the room illumination level to save energy while maintaining constant the light level perception of the users. So the energy saving is invisible to them. Our proposed ILS leverages a radiosity model to estimate the light level which is perceived by a person within an indoor environment, taking into account the person position and her/his viewing frustum (head pose). ILS may therefore dim those luminaires, which are not seen by the user, resulting in an effective energy saving, especially in large open offices (where light may otherwise be ON everywhere for a single person). To quantify the system performance, we have collected a new dataset where people wear luxmeter devices while working in office rooms. The luxmeters measure the amount of light (in Lux) reaching the people gaze, which we consider a proxy to their illumination level perception. Our initial results are promising: in a room with 8 LED luminaires, the energy consumption in a day may be reduced from 18585 to 6206 watts with ILS (currently needing 1560 watts for operations). While doing so, the drop in perceived lighting decreases by just 200 lux, a value considered negligible when the original illumination level is above 1200 lux, as is normally the case in offices.

We present the new methodology ONTO-PLC to deliver software programs on system-on-chip or single-board computers used to control industrial plants, as substitutes for programmable logic control technologies. The methodology is ontology-driven based on the abstract description of the plant at a level in which the plant itself is viewed as a set of instruments, each instrument being a set of machineries coordinated in functional terms by a control system, formed by sensors and actuators, under the control of an abstract model of behavior delivered by means of an extended finite state machine.

Recent approaches on trajectory forecasting use tracklets to predict the future positions of pedestrians exploiting Long Short Term Memory (LSTM) architectures. This paper shows that adding vislets, that is, short sequences of head pose estimations, allows to increase significantly the trajectory forecasting performance. We then propose to use vislets in a novel framework called MX-LSTM, capturing the interplay between tracklets and vislets thanks to a joint unconstrained optimization of full covariance matrices during the LSTM backpropagation. At the same time, MX-LSTM predicts the future head poses, increasing the standard capabilities of the long-term trajectory forecasting approaches. With standard head pose estimators and an attentional-based social pooling, MX-LSTM scores the new trajectory forecasting state-of-the-art in all the considered datasets (Zara01, Zara02, UCY, and TownCentre) with a dramatic margin when the pedestrians slow down, a case where most of the forecasting approaches struggle to provide an accurate solution., Comment: 10 pages, 3 figures to appear in CVPR 2018

Meteorological forecasting is the process of providing reliable prediction about the future weathear within a given interval of time. Forecasters adopt a model of reasoning that can be mapped onto an integrated conceptual framework. A forecaster essentially precesses data in advance by using some models of machine learning to extract macroscopic tendencies such as air movements, pressure, temperature, and humidity differentials measured in ways that depend upon the model, but fundamentally, as gradients. Limit values are employed to transform these tendencies in fuzzy values, and then compared to each other in order to extract indicators, and then evaluate these indicators by means of priorities based upon distance in fuzzy values. We formalise the method proposed above in a workflow of evaluation steps, and propose an architecture that implements the reasoning techniques.

We present a novel framework to directly estimate body fat percentage from depth images of human subjects and to visually evaluate salient points of the body shape related to the fat distribution. For this purpose, we created a novel, publicly available dataset including front and back depth images of a set of subjects with specific features (active young men or professional sportsmen) with associated ground truth fat values estimated with dual-energy x-ray absorptiometry (DXA) scanning. These depth images were obtained with depth rendering of an available dataset of whole body scans, simulating low-cost depth sensor acquisitions. We customized a ResNet-50 regressor to estimate fat percentage values directly from the front/back scans, achieving promising accuracy (standard errors of estimate SEE less than 2.1 on the depth renderings and 2.5 on a small set of real depth scans. We also demonstrate that, using a custom perturbation-based procedure for analyzing deep networks, it is possible to highlight, on subjects' depth images, the specific body areas related to fat accumulation (typically neck, shoulders, hip, and abdomen) and those characterizing skinny subjects (chest and abdomen).

Lexical classification is one of the most widely investigated fields in (computational) linguistic and Natural language Processing. Adjectives play a significant role both in classification tasks and in applications as sentiment analysis. In this paper a simple algorithm for lexical classification of comparable adjectives, called MORE (coMparable fORm dEtector), is proposed. The algorithm is efficient in time. The method is a specific unsupervised learning technique. Results are verified against a reference standard built from 80 manually annotated lists of adjective. The algorithm exhibits an accuracy of 76%.