Average Discovery Time Research Articles

Achievable Minimally-Contrastive Counterfactual Explanations

Decision support systems based on machine learning models should be able to help users identify opportunities and threats. Popular model-agnostic explanation models can identify factors that support various predictions, answering questions such as “What factors affect sales?” or “Why did sales decline?”, but do not highlight what a person should or could do to get a more desirable outcome. Counterfactual explanation approaches address intervention, and some even consider feasibility, but none consider their suitability for real-time applications, such as question answering. Here, we address this gap by introducing a novel model-agnostic method that provides specific, feasible changes that would impact the outcomes of a complex Black Box AI model for a given instance and assess its real-world utility by measuring its real-time performance and ability to find achievable changes. The method uses the instance of concern to generate high-precision explanations and then applies a secondary method to find achievable minimally-contrastive counterfactual explanations (AMCC) while limiting the search to modifications that satisfy domain-specific constraints. Using a widely recognized dataset, we evaluated the classification task to ascertain the frequency and time required to identify successful counterfactuals. For a 90% accurate classifier, our algorithm identified AMCC explanations in 47% of cases (38 of 81), with an average discovery time of 80 ms. These findings verify the algorithm’s efficiency in swiftly producing AMCC explanations, suitable for real-time systems. The AMCC method enhances the transparency of Black Box AI models, aiding individuals in evaluating remedial strategies or assessing potential outcomes.

A search method for a hypersonic gliding vehicle based on early warning information guidance

The high speed and high manoeuvrability of a hypersonic gliding vehicle (HGV) pose a severe challenge to the existing early warning detection system. Optimising the search resources of a phased array radar to improve the HGV detecting efficiency has become a practical problem. In this study, a HGV search method based on a hybrid optimisation algorithm is proposed with the early warning information guidance as a priori condition. Firstly, the HGV priority judgement model is established, and the priority quantification equations for height, velocity and distance are designed. Secondly, the search parameters of the radar are optimised with the maximum cumulative detection probability, the shortest average discovery time and the highest priority level as the objective functions, and a radar search model is established. Finally, to overcome the problem that particle swarm optimisation (PSO) is easy to fall into local optimal solution, a hybrid optimisation algorithm based on differential evolution (DE) and PSO is proposed. The performance of the proposed method is verified in two simulation scenarios, and the results show that the proposed method outperforms existing mainstream search methods and can reasonably allocate search resources according to the priority of HGV.

A Dominating Set Routing Scheme for Adaptive Caching in Ad Hoc Network

Current efforts for providing an efficient dynamic source routing protocol (DSR) for use in multi-hop ad-hoc wireless are promising. This is since DSR has a unique characteristic in that it uses source routing, instead of relying on the routing table at each intermediate device. This study addresses the current challenges facing DSR protocol in terms of the dynamic changes of the route and how to update such changes into the route cache of the DSR. The challenges typically persist when a sudden route break occurs resulting in a delay in updating the new node location into the cache of the DSR protocol. For that, this study proposes a novel algorithm to improve the cache updating of DSR protocol in the ad-hoc network using dominating set-based routing (DBR). In DBR, the dominating nodes establish node update cache in accordance with the characteristics of the new route. Network Simulator version 2 (NS2) was used to implement and evaluate the proposed algorithm. A comparison of certain performance metrics was carried among the proposed DSR-DBR, DSR-route-cache, and DSR-original in the transmission control protocol and user datagram protocol. The DSR-DBR performance result showed a significant improvement in the average throughput, average end-to-end delay, average discovery time and routing overhead.

Beam-Sweeping Design Based on Nearest Users Position and Beam in 5G mmWave Networks

The beam sweeping procedure estimates the beamforming directions, or beams, to be used for downlink millimeter-wave cellular transmissions to an incoming user entering the cell. We propose three new approaches for the choice of the sequence by which beams are explored (the beam-search sequence), in order to reduce the average time to find the directions. The proposed methods exploit a) the correlation among angles of departure of the incoming user and its nearest, already connected user, and b) the position of both the incoming and its nearest user. For point a), the nearest-neighbor-beam search method starts from the beam of the nearest user, and then explores other beams, with directions progressively farther away (in the angular domain) from the first beam. In the beam-and-angle search, the beam-search sequence is optimized by exploiting the knowledge of both users' position and the nearest-user beam. We also propose the simplified beam-and-angle search, which still exploits the nearest-user information, although in a suboptimal and simpler fashion. Based on the channel model of the 3rd generation partnership project, we analytically design the optimal beam-and-angle search sequence. We assess the performance of the proposed solutions, in terms of their average discovery time, also in comparison with existing approaches.

Fast Initial Access for mmWave 5G Systems with Hybrid Beamforming Using Online Statistics Learning

In cellular communication systems, user discovery or initial access (IA) procedures enable a terminal entering a cell to connect to the base station (BS). Up to 4G systems, users broadcast a cell-connect request omnidirectionally, and the BS replies to the request. In 5G systems operating at mmWave, directional transmissions (by means of beamforming) are needed to overcome strong attenuations. However, during IA the user's and BS's reciprocal positions are not known, and therefore a search procedure explores different directions, substantially increasing the IA overhead. Still, directions of arrival statistics largely depend on buildings and typical user entrance paths, which characterize the cell. By exploiting these statistics IA can be significantly accelerated, by searching more often where users typically are. In order to acquire the statistics, an online method is proposed and its performance is assessed for the first time using real data from a traffic data survey. We observe that the online method reduces the average discovery time up to 25 percent with respect to the conventional approach that does not exploit the user statistics.

Statistical approaches for initial access in mmWave 5G systems

AbstractWhen a new user enters a cell in a mmWave cellular system, the beamforming directions must be identified to initiate communication, a procedure known as initial access (IA). However, users are more likely to enter from some directions than others (eg, along streets), and beamforming directions (eg, those affected by blockage) may never be used. We exploit the unequal distribution of the entrance direction to speed up the IA procedure, exploring more often the directions wherein the probability of finding new users is higher. Two solutions are proposed: a memory‐less random illumination (MLRI) algorithm and a statistical and memory‐based illumination (SMBI) algorithm. While in MLRI, the direction to be explored is randomly generated according to an optimized distribution, and independent of previously explored directions, in SMBI a precise exploration sequence is designed, thus we take into account previously explored directions. In the analysis, we include the movement of the user within the cell during the IA process, described by a Markov chain whose states correspond to the beamforming directions associated to the user position at a given IA exploration time. We assess the performance of the proposed methods in terms of average discovery time.

On Expected Neighbor Discovery Time With Prior Information: Modeling, Bounds and Optimization

Neighbor discovery (ND) is an essential prerequisite for any peer-to-peer communication. In general, minimizing the discovery time is the goal for ND schemes. In this paper, we study the average discovery time for directional random ND when nodes have prior information about their set of possible neighbors, which also helps identify the performance limits of random ND schemes. Typically, discovery time analysis is done for assumptions that simplify the network structure, such as uniform neighbor relations for all nodes. However, with prior information the directional transmission probabilities depend on the node and the direction. This complicates the analysis of the expected discovery time. We first provide a closed-form expression for the expected discovery time based on the non-uniform coupon collector problem. Next, we identify directional transmission probabilities of each node that achieve a small discovery time. Due to the mathematical complexity, we provide a lower and an upper bound on the expected discovery time, which allows us to write the problem as a convex optimization problem. Through simulations, we demonstrate the performance gain due to prior knowledge with the proposed methods as compared with when no prior information is available, as well as the impact of uncertainty in the prior knowledge.

Improving passive discovery for IEEE 802.15.4-based mobile body sensor networks

We investigate passive discovery of IEEE 802.15.4-based body sensor networks (BSNs). BSNs are wearable networks that monitor the vital functions of the body while the person can roam around freely. One challenging task that BSNs have to perform is the discovery of certain coordinator nodes, for example, gateways towards data processing and storage services. We are presenting a cooperative passive discovery scheme (the rumour-based scheme) for the beacon-enabled mode. We consider a simple scenario where a specific mobile BSN is searching for a specific destination network. We evaluate the performance of our scheme in terms of the average discovery time and the discovery probability using simulations. We also assess the influence of some important system parameters, e.g. the speed of mobile BSNs, the beacon order, the mobility model or the number of nodes used for discovery. The results show that our strategy can significantly reduce the time required to discover the target network.

Evaluating visual attention for multi-screen television: measures, toolkit, and experimental findings

New and emerging multi-screen television scenarios and applications need new evaluation procedures, methodologies, and tools to support multi-screen data analysis. In this work, we introduce a set of 12 measures to characterize viewers’ visual attention patterns for multi-screen TV. Out of these, eight measures are computed directly from eye-tracking data, while the other four are evaluated using questionnaires. We applied our measures during a controlled experiment involving nine distinct screen layouts with two, three, and four TV screens, for which we report new findings about viewers’ distributions of visual attention. For example, we found that viewers need an average discovery time up to 4.5 s to visually fixate four screens, and their subjective perceptions of what they watched and for how long they watched each screen are substantially accurate, i.e., we report Pearson’s correlation coefficients up to r = .892 with ground truth measured with eye-tracking equipment. We also analyze and discuss the evolution of our participants’ distributions of visual attention over time from the perspective of our new set of measures. For example, we found that people perform significantly more transitions between screens during the first seconds of watching television, after which their level of visual attention converges to a stable value. We complement the findings revealed by our objective eye gaze measurements with subjective data about participants’ perceived cognitive work load and comfortability while watching more than one TV screen, and we measure viewers’ capacities to understand and recall content delivered simultaneously on multiple screens. To foster new studies and explorations of viewers’ visual attention patterns during multi-screen television watching, we release in the community an update for an existing software toolkit (i.e., VATic-TV, the Visual Attention Toolkit for TV, now at version v2) that automatically computes our measures from data delivered by standard eye-tracking equipment. We hope that our new set of measures and the companion software will benefit the community as a first step toward understanding visual attention for emerging multi-screen TV applications and, consequently, will help researchers and practitioners to design new TV applications that will better exploit viewers’ visual attention patterns toward new, richer television experiences.

Improving safety in the operating room: a systematic literature review of retained surgical sponges

Gossypibomas are surgical sponges that are unintentionally left inside a patient during a surgical procedure. To improve this patient safety indicator, anesthesiologists will need to work with operating room personnel. This study's goal was to systematically review the literature on retained sponges to identify body location, time to discovery, methods for detection, and risk factors. Two hundred and fifty-four gossypiboma cases (147 reports from the period 1963-2008) were identified via the National Library of Medicine's Medline and the Cochrane Library. Gossypibomas (mean patient age 49 years, range 6-92 years) were most commonly found in the abdomen (56%), pelvis (18%), and thorax (11%). Average discovery time equaled 6.9 years (SD 10.2 years) with a median (quartiles) of 2.2 years (0.3-8.4 years). The most common detection methods were computed tomography (61%), radiography (35%), and ultrasound (34%). Pain/irritation (42%), palpable mass (27%), and fever (12%) were the leading signs and symptoms, but 6% of cases were asymptomatic. Complications included adhesion (31%), abscess (24%), and fistula (20%). Risk factors were case specific (e.g. emergency) or related to the surgical environment (e.g. poor communication). Most gossypibomas occurred when the sponge count was falsely pronounced correct at the end of surgery. More is being discovered about the patterns leading to a retained sponge. Multidisciplinary approaches and new technologies may help reduce this low frequency but clinically significant event. However, given the complexity of surgical care, eliminating retained sponges may prove elusive.

Adaptive energy conserving algorithms for neighbor discovery in opportunistic Bluetooth networks

In this paper, we introduce and evaluate novel adaptive schemes for neighbor discovery in Bluetooth-enabled ad-hoc networks. In an ad-hoc peer-to-peer setting, neighbor search is a continuous, hence battery draining process. In order to save energy when the device is unlikely to encounter a neighbor, we adaptively choose parameter settings depending on a mobility context to decrease the expected power consumption of Bluetooth-enabled devices. For this purpose, we first determine the mean discovery time and power consumption values for In different Bluetooth parameter settings through a comprehensive exploration of the parameter space by means of simulation validated by experiments on real devices. The fastest average discovery time obtained is 0.2 s, while at an average discovery time of I s the power consumption is just 1.5 times that of the idle mode on our devices. We then introduce two adaptive algorithms for dynamically adjusting the Bluetooth parameters based on past perceived activity in the ad-hoc network. Both adaptive schemes for selecting the discovery mode are based only on locally-available information. We evaluate these algorithms in a node mobility simulation. Our adaptive algorithms reduce energy consumption by 50% and have up to 8% better performance over a static power-con serving scheme

Discovery of selected water dispensers by newborn pigs

Newborn pigs in 86 litters were offered drinking water from four types of dispenser to determine how quickly each design would be discovered. On average piglets discovered water within about 24 h from an exposed water surface (bowl or cup), whereas discovery time was delayed to more than 72 h with nipple or push-lever dispensers. A prototype dispenser with a wide bowl and continuous bubbling action reduced (P < 0.05) average discovery time to about 14 h. Key words: Piglet, water, dispenser design