In this paper, we bring a human in the loop, and enable a human teacher to give feedback to a keyword extraction framework in the form of natural language. We focus on the problem of keyphrase extraction in which the quality of the output can easily be judged by non-experts.

With the ubiquitous nature of WiFi devices and WiFi networks around us, which continuously transmit and receive high bandwidth data, these transmission channels can be used as an effective sensing medium for detecting movements. With the use of Machine Learning, we develop a model using the features of the channel state information to train and predict these movements accurately. This paper proposes a demonstration harnessing the low-cost WiFi-based IoT for the detection of movements. Such a setup will facilitate the usage of movements like hand gestures to carry out common tasks such as changing the television channel, changing slides during a presentation, no-controller wireless gaming, etc. We achieve an F1 score of 0.99 with the latency of 45.75ms.

As machine learning algorithms make further inroads into society, calls are increasing from multiple stakeholders for these algorithms to explain their outputs. Moreover, these stakeholders, whether they be government regulators, affected citizens, domain experts, or developers, present different requirements for explanations. To address these needs, we introduce AI Explainability 360, an open-source software toolkit featuring eight diverse state-of-the-art explainability methods, two evaluation metrics, and an extensible software architecture that organizes these methods according to their use in the AI modeling pipeline. Additionally, we have implemented several enhancements to bring research innovations closer to consumers of explanations, ranging from simplified, more accessible versions of algorithms, to guidance material to help users navigate the space of explanations along with tutorials and an interactive web demo to introduce AI explainability to practitioners. Together, our toolkit can help improve transparency of machine learning models and provides a platform to integrate new explainability techniques as they are developed.

The Internet provides a platform for various commercial activities and it is important to ensure that it remains a level playing field for all the players. Several countries have enacted laws such that the Internet remains neutral by prohibiting preferential treatment of traffic of one application or content over the other. However to enforce such regulations one needs to detect any violations. In this work we demonstrate a method to identify non-neutral behavior by comparing the quality of service received by different applications `when they experience similar network conditions.'

Mixed reality is one of the trending technology. It provides an excellent platform to the users, as it includes good graphics and animations. Mixed reality provides many applications in gaming, education, healthcare etc. This domain, can be considered for medical purposes i.e., rhinoplasty, which is a surgical procedure used to correct the nose and other parts of the face. Sometimes the result obtained after the surgery, is not convincing. This paper illustrates the method to overcome this problem by using Mixed Reality Toolkit(MRTK) over Unity gaming platform. Therefore the patient and the doctor can have discussion and edit according to the patient's convenience and further proceed to surgery.

"We describe N V iZ a graphical visualization tool developed to visualize log data generated by network security devices and services like firewall, intrusion detection system and domain name system. It can generate interactive visualization graphs identifying popular websites visited, active users, DNS query patterns, IDS alert types and their distribution and also peer network connections of infected and non-infected machines."

Industry 4.0 is the latest trend in the manufacturing sector that focuses on intelligent manufacturing and smart factories. This leads to opportunities in automation, optimization, asset management and predictive maintenance, which helps reduce downtime and increase revenue. In this paper, we propose the solution that was created for a particular SMT PCB manufacturing facility in Mysuru (Vinyas IT), the features it has to offer and the methodologies that were implemented in order to achieve our goals. We will also highlight the important aspects of the solution that will be showcased during the demonstration and the impact of our solution.

In the recent times, smartphones or cell phones have become the most widely used for communication. This application mainly aimed to provide a simplified management of prohibitions enforced in public places by developing an application to control the different mobile features. There are many places which includes hospital, religious places, research centers, museums, corporate offices, petrol pumps, universities etc., where they clearly mentioned that, "usage of mobile phones is prohibited", "keep your mobile phones silent", "usage of camera is restricted", etc. many times people forget to switch off their mobile phone or forget to switch their mobile phone to silent or vibrate mode, which is not feasible. Capturing of photos and videos even in the places where the mobile phones usage is restricted. In this paper, we develop an application that provides complete control on automated profile switching and camera disabling of mobile phones.

Visible Light Communication (VLC), which operates in the Terahertz band, is theoretically capable of terabits/second data speeds. However, traditional state-of-the-art VLC has largely been limited in the data rates, which also trade off with communication range. A common approach to modulation in VLC is the use of ON-OFF Keying (OOK), where a binary bit 1 is mapped to a high or ON state of the light emitter and bit 0 to a low or OFF state. While there have been approaches to improve the spectral efficiency through other modulation schemes such as color-shift-keying (CSK), frequency-shift-keying (FSK) and orthogonal frequency division multiplexing (OFDM), the fundamental mapping of intensity of light emitters to high and low states is still required. In this work, we explore a novel approach for using multi-level or M-ary amplitude modulation to encode symbols in VLC. We demonstrate a technique that uses machine learning of the VLC channel state to improve demodulation of M-ary signals in a LED-Photodiode VLC system.

"Parking in cities is regulated by policies prepared by the city council, and are designed to maximize service while keeping sufficient road space for unobstructed traffic flow. Keeping this in view, enforcement agencies mandate vehicles to not park at locations that are marked with no-parking signages. However, in developing regions of the world, a weak compliance to this policy is often observed.
StreetHAWK is the first-of-its-kind real-time, automated, city-scale system that addresses this problem using visual sensing and edge computing models on readily available smartphones. The system overcomes the challenges of small object detection, limited visual ranging (with embedded single camera phone units), and other sources of detection inaccuracies to identify (on an average) 80% of on-street parking violations compared to manual approaches."

While 5G systems are being deployed across the globe in a phased manner, trust models for communication between the User Equipment (UE) and the core network play a vital role in fulfilling the network goals promised by 5G. Towards that end, the 5G - Authentication and Key Agreement (5G-AKA) protocol has been standardised by the 3GPP consortium after multiple iterations and corrective responses. Despite this progress, there is a requirement to collate information from multiple ETSI technical specifications and several research contributions to implement a standardised security architectural model for 5G networks. Identifying the absence of single point repository containing 5G security protocols, this demonstration provides first hand collation and description of state-of-the-art details in simple and unified manner. In this demonstration, we implement the 3GPP compliant 5G-AKA protocol on a network of XBee S2C devices wherein the protocol involves a sequence of modules involving secure authentication, key exchange and payload transmission. To implement the security protocol, we extract the precise recommendations of 3GPP and also use open source algorithms wherever the implementation is left open as proprietary solution. Other than implementing standard modules from 3GPP specifications, we also showcase novel key exchange and provenance-recovery mechanisms that potentially fit the 5G standards in the framework of device-to-device (D2D) communication.

"In a smart IoT system, multi-sensing at a field node is a typical scenario. The examples considered in this study are pollution monitoring and smart energy metering. In such applications, energy sustainability and communication and storage resource usage optimization are two of the key issues of interest. In this study, on one hand it is intended to develop indigenous beyond state of the art multi-sensing boards with the inherent smartness in energy replenishment and sensing/communication activities. On the other hand, smart data collection and processing at the end node (fog node or edge node) is of interest primarily from efficient communication bandwidth usage perspective.
On the first exercise towards energy sustainable IoT sensing and communication board design, we have designed a prototype for a 5G capable environmental air pollution monitoring system. The system measures concentrations of NO2, ozone, CO and SO2 using semiconductor sensors. Further, the system gathers other environmental parameters like temperature, humidity, PM1, PM2.5 and PM10. The prototype is equipped with a GPS sub-system for accurate geo-tagging. The board communicates through Wi-Fi and NB-IoT. The board is also equipped with energy harvesting power management, and is powered through solar energy and battery backup.
On the second exercise, a working model of a smart IoT device with a data pruning subsystem is designed, where a smart energy meter is considered for an example application. As a proof of concept we plan to demonstrate data compression at the edge to save bandwidth required for data transmission to a remote cloud. At each smart meter, sparsity of data is exploited to devise an adaptive data reduction algorithm using compressive sampling technique such that the bandwidth requirement for smart meter data transmission is reduced with minimum loss of information. The Smart Energy Meter is WiFi and NB-IoT enabled. This meter is capable of logging multiple energy consumption parameters.
The overall objective has been demonstrating the ability of beyond state of the art circuits and system design for IoT communications, wherein context specific intelligence is applied at the at the node. The broad philosophy in this study can be readily extended to any chosen IoT application."