Recent research has identified intertemporal impulsivity as a critical cognitive variable for explaining the autocatalytic nature of socioeconomic status. But how exactly this relationship transpires has not been clearly identified, with several possible cognitive mechanisms having been proposed in the literature. We designed an experimental paradigm in which participants farmed crops under budgetary constraints and intermittently faced random resource demands. We discovered that, as a result of unpredictable resource shocks, people's preferences shifted from long-term choices to short-term ones. We also found people's self-reported sense-of-control scores to be predictive of the magnitude of their preference shifts. On the basis of these results, we argue that steep inter-temporal discounting arises as a rational adaptation to persistently experiencing long-term planning failures due to unpredictable resource shocks.


Unpredictability shortens planning horizons

We consider partially observable Markov decision processes (POMDPs) modeling an agent that needs a supply of a certain resource (e.g. electricity stored in batteries) to operate safely. The resource is consumed by agent's action and can be replenished only in certain states. The agent aims to minimize the expected cost of reaching some goal while preventing resource exhaustion, a problem we call resource-safe goal optimization (RSGO). We take a two-step approach to the RSGO problem. First, using formal methods techniques, we design an algorithm computing a shield for a given scenario: a procedure that observes the agent and prevents it from using unsafe actions that might eventually lead to resource exhaustion. Second, we augment the POMCP heuristic search algorithm for POMDP planning with our shields to obtain an algorithm solving the RSGO problem. We implement our algorithm and present experiments showing its applicability to benchmarks from the literature.

Shielding in Resource-Constrained Goal POMDPs

The Internet of Vehicles (IoV) has been recognized as a powerful application of the Internet of Things (IoT) in the Intelligent Transportation System (ITS), providing intelligence for interconnection between devices, interaction with the environment, and thus, greater efficiency in sensor data exploitation. Therefore, leveraging the huge capability of sensors embedded in vehicles to offer a Sensing As A Service (Se-aaS) represents a great solution to exploit under-used sensor resources and continue providing sensors despite their positions and mobility patterns.
Nevertheless, the high network mobility and the fast topology changes in IoV impact the vehicle availability and complicate the service provision. To this aim, we propose a vehicle sensor virtualization in a Cloud IoV architecture that encompasses functional blocks of mobile sensor suppliers, Sensor Cloud Service Provider (SCSP), and service consumers. Moreover, we propose a reinforcement learning-based solution for vehicle sensor selection to predict and dynamically select the physical sensors composing the vehicle virtual sensor. The conducted simulations show the effectiveness of the proposed solution.

Reinforcement Learning-based Virtual Sensor Provision in Internet of Vehicles (IoV)

It becomes very common to use cell phones in public transportation and the cars. Vehicular networking has a major problem which is the degradation of signal quality due to interference and the large number of mobile devices. Artificial intelligence (AI) is a promising technique for next-generation wireless networks. Deep learning is a type of AI derived from machine learning; here the machine can learn by itself, unlike programming where it is content to execute rules to the letter predetermined. In addition, AI can be explored in order to solve various problems. In this paper, we tackle the problem of resource allocation in a vehicular small cell network (VSCN). Indeed, we propose a new mechanism based on deep reinforcement learning denoted Resource Allocation based Deep Reinforcement Learning (RA-DRL). The main goal of our proposed method is to maximize the total system sum rate (throughput) while guaranteeing minimum interferences, Quality of Service (QoS) and the demand for all users. Simulation results demonstrate that our proposed RA-DRL algorithm exhibits better performance comparing to the other methods, by maximizing the total system sum rate while maintaining inter-VSCs interferences and a minimum latency

Deep Reinforcement Learning for Downlink Resource Allocation in Vehicular Small Cell Networks

Under the vision of industry 4.0, industrial networks are expected to accommodate a large amount of aggregated traffic of both operation and information technologies to enable the integration of innovative services and new applications. In this respect, guaranteeing the uninterrupted operation of the installed systems is an indisputable condition for network management. Network measurement and performance monitoring of the underlying communication states can provide invaluable insight for safeguarding the system performance by estimating required and available resources for flexible integration without risking network interruption or degrading network performance. In this work, we propose a data-driven in-band telemetry method to monitor the aggregated traffic of the network at the switch level. The method learns and models the communication states by local network-level measurement of communication intensity. The approximated model parameters provide information for network management for prognostic purposes and congestion avoidance resource planning when integrating new applications. Applying the method also addresses the consequence of telemetry data overhead on QoS since the transmission of telemetry packets can be done based on the current state of the network. The monitoring at the switch level is a step towards the Network-AI for future industrial networks. 

Data driven method for In band Network Telemetry Monitoring of Aggregated Traffic

Industrial Internet-of-Things (IoT) massively deploys intelligent computing in industrial production and manufacturing environments seeking automation, reliability, and control. Machine Learning models provide intelligent decisions to drive manufacturing systems to the next level of productivity, efficiency, and safety. One of the critical challenges that must be faced is the deployment of Machine Learning models at the network edge to detect data anomalies caused by Industrial IoT hardware failures, since industrial IoT devices are prone to errors and failures. These anomalies can harm the industrial IoT system by producing false alarms, consuming network resources, and affecting productivity. Because of that, it is critical to rely on low latency and high precision detection systems to verify the data received from industrial IoT devices. In light of this, we assessed key performance indicators of five machine learning models running at edge computing, to provide in-depth discussions. The performance results were obtained from an oil refinery scenario using a real industrial IoT dataset. The performance was measured in terms of (a) Accuracy, (b) Precision, (c) Recall, (d) F1 score, (e) Training time, and (f) Response time.

Underline digital video library

Unpredictability shortens planning horizons

Shielding in Resource-Constrained Goal POMDPs

Reinforcement Learning-based Virtual Sensor Provision in Internet of Vehicles (IoV)

Deep Reinforcement Learning for Downlink Resource Allocation in Vehicular Small Cell Networks

Data driven method for In band Network Telemetry Monitoring of Aggregated Traffic

On the Performance of Machine Learning at the Network Edge to Detect Industrial IoT Faults

Performance Improvement and Reliability Physics in SiC MOSFETs

New Reliability Ecosystem: Maximize Technology Value to Serve Diverse Markets

Silicon Lifecycle Management: Increasing Semiconductor Predictability in an Unpredictable World

Transformation of Memory and Storage Hierarchy: Enabling high performance, reliable, and sustainable compute applications

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES