Harshvardhan J. Pandit


Currently a Ph.D researcher with ADAPT research center at School of Computer Science and Statistics of Trinity College Dublin. Started in April 2016 and expected to graduate by April 2020. See my work at OpenScience - ADAPT.
Google Scholar
DBLP search

PhD Research

To develop an appropriate design for a privacy reasoning service (primarily within the browser) that can consume open data usage declarations for service providers, and context-aware privacy setting from the user in order to advise users on the compatibility or risk associated with use of specific service features. It will advise users on the release of personal service interaction data and contextual meta-data e.g. collected from device GPS, motion sensors, speech recognition and video processing, to requesting service providers, based on the outcome of semi-automated privacy-aware decisions (in collaboration with other researchers. This may also be extended to open processing platforms in a user’s home areas network, where different privacy preferences, sensed context and service usage from multiple users must be considered. The performance envelope for different degrees of reasoning for advice generation will be assessed.


Previous Projects

A Model for Contextual Data Sharing in Smartphone Applications

Masters in Computer Science by Research at University College Cork, Ireland in Oct 2016.

This research investigated how smartphone apps can perform contextual data sharing with an emphasis on practical integration into the existing platforms and app models. The identification of information and its associated context is necessary to create context definitions that allow different apps to identify the context of the shared data. An approach to model the Context Definitions using computer science concepts such as object-oriented data structures provides flexibility. A context datastore is defined to store and share contextual information between apps, which creates an independence between apps for acquiring information and provides compatibility with the existing security models on various platforms. The model allows apps to retrieve contextual data in a simple and efficient manner without interacting directly with the other apps. read more...

KWEST: A Semantically Tagged Virtual File System
Bachelors of Engineering in Computers from Pune University, India in September 2014.
final-year group project with Aseem Gogte, Rohit Sharma, Sahil Gupta
The limitation of data representation in traditional file systems is that data representation is bound only in a single way of hierarchically organising files. A semantic file system provides addressing and querying based on the content rather than storage location. Semantic tagging is a new way to organise files by using tags in place of directories. In traditional file systems, symbolic links become non-existent when file paths are changed. Assigning multiple tags to each file ensures that the file is linked to several virtual directories based on its content. By providing semantic access to information, users can organise files in a more intuitive way. In this way, the same file can be accessed through more than one virtual directory. The metadata and linkages for tagging are stored in a relational database which is invisible to the user. This allows efficient searching based on context rather than keywords. The classification of files into various ontologies can be done by the user manually or through automated rules. For certain files types, tags can be suggested by analysing the contents of files. The system would be modular in design to allow customisation while retaining a flexible and stable structure. read more...

Recent Publications

see all publications

Modelling provenance for GDPR compliance using linked open data vocabularies
(2017 Workshop) Harshvardhan J. Pandit, Dave Lewis. Society, Privacy and the Semantic Web - Policy and Technology (PrivOn), co-located with ISWC 2017
proceedings online published (PDF) alternate PDF download

Compliance through Informed Consent: Semantic Based Consent Permission and Data Management Model
(2017 Workshop) Kaniz Fatema, Ensar Hadziselimovic, Harshvardhan J. Pandit, Dave Lewis. Society, Privacy and the Semantic Web - Policy and Technology (PrivOn), co-located with ISWC 2017
proceedings online published (PDF) alternate PDF download

Linked Data Contracts to Support Data Protection and Data Ethics in the Sharing of Scientific Data
(2017 Workshop) Ensar Hadziselimovic, Kaniz Fatema, Harshvardhan J. Pandit, Dave Lewis. Sharing of Scientific Data in Enabling Open Semantic Science (SemSci), co-located with ISWC 2017.
proceedings online published (PDF) alternate PDF download