Akshit Singhal

Two Families of Optimal Multipath Congestion Control Protocols

In this work, we provide motivation for a zero-effort crowdsensing task: auto-annotated ground tr... more In this work, we provide motivation for a zero-effort crowdsensing task: auto-annotated ground truth collection for physical activity recognition. Data obtained through Smartphones for classification of human activities is prone to discrepancies, which reiterates the need for better and larger activity datasets. Artificial data generation algorithms fail to efficiently generate quality instances for minority data. In the proposed model, crowd-sourced sensor data is classified by a robust classifier built by researchers ground up. We nominate a Generic Classifier with ≥ 95% accuracy for this purpose. Data collection and distribution models which ensure that the crowd client receives non-skewed, quality data from locations with higher degree of activity occurrence are elucidated upon. Also integrated within our proposed model are Location-Specific Classifiers, which can be utilized by developers to optimize on location-specific tasks. Effective validation of classified activities using diverse sensor data streams improves the proposed classifier systems and boosts ground-truth accuracy.

Smartphone-based Qualitative Analyses of Social Activities During Family Time

With the evolution of nuclear families and diverse career options, families as social groups are ... more With the evolution of nuclear families and diverse career options, families as social groups are spending lesser time together than in the past decades. In this work, we study both quantitative as well as qualitative aspects of time spent with family members through a smartphone-based pervasive study on a sample of 12 families over 14 days. Further, we also examine the perception of 78 millennials on what they feel about, and expect from, the time they spend with their families, however long it may be. We aim to identify the key parameters that shape family life in this day and age, along with examining the participation of individuals of various roles within the family in activities such as conversations, workout sessions, eating together and other social interactions. Among all activities detected to be performed by families reporting high satisfaction with familial life, Eating Together and Using Smartphones Together emerged as the most prominent ones. We discover a greater disparity among the habits of family members, especially millennials, staying away from each other as compared to those staying together.

PeopleSave: Recommending effective drugs through web crowdsourcing

In this paper, we describe PeopleSave-a drug recommendation and feedback system for doctors on th... more In this paper, we describe PeopleSave-a drug recommendation and feedback system for doctors on the basis of contextual patient reviews crowd-sourced from the Internet. Unlike other systems proposed in the past, we filter information sources to check for crowdsourcing feasibility and then assess the drug's effectiveness based on its reported detrimental effect on a patient. This helps in eliminating certain drugs that would almost certainly have an adverse effect on the patient's health and thereby obtain a set of recommendable drugs. These recommendations are further refined by analyzing the sentiment behind the opinions of patients who have been administered these drugs in the past. The resultant set of prescribable drugs agrees with those suggested by the consulted physicians for the considered sample set of diabetes patients. The critical assessment of the prototype system for Diabetes Type II drugs by both doctors and patients also reiterates the need for a feedback system that can possibly go a long way in improving patient experience of a drug. This leads us to conclude that PeopleSave, as a combination of the recommendation system prototype and the proposed feedback system, can be successful in improving the process of prescription of medicines for a varied range of medical conditions.

Download

Two Families of Optimal Multipath Congestion Control Protocols

2022 IEEE 30th International Conference on Network Protocols (ICNP)

HOLNET: A Holistic Traffic Control Framework for Datacenter Networks

2020 IEEE 28th International Conference on Network Protocols (ICNP), 2020

In this paper, we put forward a HOListic traffic control framework for datacenter NETworks (HOLNE... more In this paper, we put forward a HOListic traffic control framework for datacenter NETworks (HOLNET). HOLNET reformulates the network utility maximization (NUM) framework into a HOLNET NUM framework that fully harnesses the potential of the existing NUM-based solutions to allow large families of traffic control protocols of various degrees of sophistication to be developed, i.e., host-based, single or multiple Class-of-Service (CoS) enabled, single or multi-path congestion control, with or without in-network load balancing. Unlike the existing solutions that are largely empirical and point by design, HOLNET is a principled, systematic framework. All the protocols in a family developed under HOLNET share a common, userdefined global optimization objective and fairness criterion. As a result, the protocols in a family can be fairly compared and carefully selected to fully explore the performance, scalability and design complexity tradeoffs. Case studies, based on both a single and a multi-path host-based solutions, demonstrate the viability and flexibility in HOLNET design space exploration. To further test the backward compatibility and performance with respect to some existing lightweight solutions, we develop HOLNET-UTA, an integrated congestion control and load balancing protocol, achieving TCP-fair resource allocation. HOLNET-UTA is found by simulation to improve the average flow completion time (FCT) by more than 20%, compared to DRILL with DCTCP.

Download

Smartphone-based Qualitative Analyses of Social Activities During Family Time

Proceedings of the First International Workshop on Human-centered Sensing, Networking, and Systems, 2017

With the evolution of nuclear families and diverse career options, families as social groups are ... more With the evolution of nuclear families and diverse career options, families as social groups are spending lesser time together than in the past decades. In this work, we study both quantitative as well as qualitative aspects of time spent with family members through a smartphone-based pervasive study on a sample of 12 families over 14 days. Further, we also examine the perception of 78 millennials on what they feel about, and expect from, the time they spend with their families, however long it may be. We aim to identify the key parameters that shape family life in this day and age, along with examining the participation of individuals of various roles within the family in activities such as conversations, workout sessions, eating together and other social interactions. Among all activities detected to be performed by families reporting high satisfaction with familial life, Eating Together and Using Smartphones Together emerged as the most prominent ones. We discover a greater disparity among the habits of family members, especially millennials, staying away from each other as compared to those staying together.

AnnoTainted

Proceedings of the 3rd International on Workshop on Physical Analytics, 2016

In this work, we provide motivation for a zero-effort crowdsensing task: auto-annotated ground tr... more In this work, we provide motivation for a zero-effort crowdsensing task: auto-annotated ground truth collection for physical activity recognition. Data obtained through Smartphones for classification of human activities is prone to discrepancies, which reiterates the need for better and larger activity datasets. Artificial data generation algorithms fail to efficiently generate quality instances for minority data. In the proposed model, crowd-sourced sensor data is classified by a robust classifier built by researchers ground up. We nominate a Generic Classifier with ≥ 95% accuracy for this purpose. Data collection and distribution models which ensure that the crowd client receives non-skewed, quality data from locations with higher degree of activity occurrence are elucidated upon. Also integrated within our proposed model are Location-Specific Classifiers, which can be utilized by developers to optimize on location-specific tasks. Effective validation of classified activities using diverse sensor data streams improves the proposed classifier systems and boosts ground-truth accuracy.

PeopleSave: Recommending effective drugs through web crowdsourcing

2016 8th International Conference on Communication Systems and Networks (COMSNETS), 2016

In this paper, we describe PeopleSave-a drug recommendation and feedback system for doctors on th... more In this paper, we describe PeopleSave-a drug recommendation and feedback system for doctors on the basis of contextual patient reviews crowd-sourced from the Internet. Unlike other systems proposed in the past, we filter information sources to check for crowdsourcing feasibility and then assess the drug's effectiveness based on its reported detrimental effect on a patient. This helps in eliminating certain drugs that would almost certainly have an adverse effect on the patient's health and thereby obtain a set of recommendable drugs. These recommendations are further refined by analyzing the sentiment behind the opinions of patients who have been administered these drugs in the past. The resultant set of prescribable drugs agrees with those suggested by the consulted physicians for the considered sample set of diabetes patients. The critical assessment of the prototype system for Diabetes Type II drugs by both doctors and patients also reiterates the need for a feedback system that can possibly go a long way in improving patient experience of a drug. This leads us to conclude that PeopleSave, as a combination of the recommendation system prototype and the proposed feedback system, can be successful in improving the process of prescription of medicines for a varied range of medical conditions.

Download

Tail Latency Prediction for Datacenter Applications in Consolidated Environments

2019 International Conference on Computing, Networking and Communications (ICNC)

Consolidating applications is a practical necessity in today's datacenters to reduce cost and imp... more Consolidating applications is a practical necessity in today's datacenters to reduce cost and improve resource utilization. However, resource sharing among different applications may result in high latency in responses to user requests. Due to the lack of a performance model for tail latency of Fork-Join structures, which underlay the workflows of lots of datacenter applications, the current practice is to overprovision resource in an attempt to satisfy as many user requests as possible. However, this practice leads to low resource utilization. Therefore, it is of importance to have a performance model that can accurately predict tail latency in such an environment, especially at high load regions, where resource provisioning is desired at most. In this paper, we propose an analytical solution for the prediction of tail latency of a target application in a consolidated environment where it is mixed with other background applications. The proposed model is validated against simulation through extensive case studies. The experimental results show the effectiveness of the proposed model in tail latency prediction at high load region, yielding all the prediction errors well within 10% at the load of 75% or higher, making the model a valuable tool for resource provisioning and supporting scheduling decisions in datacenter clusters to guarantee user satisfactions.

Download

Uploads

Papers by Akshit Singhal

Log In