Select a calendar:
Filter October Events by Event Type:
SUNMONTUEWEDTHUFRISAT
University Calendar
Events for October
-
PhD Defense - Anil Ramakrishna
Wed, Oct 02, 2019 @ 12:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
University Calendar
PhD Candidate: Anil Ramakrishna
Committee:
Shri Narayanan (chair)
Aiichiro Nakano
Morteza Dehghani
Location: RTH 320
Time: October 2nd, 12 pm.
Title: Computational Models for Multidimensional Annotations of Affect
Abstract: Affect is an integral aspect of human psychology, it acts as a regulator for all our interactions with external stimuli. Affect includes several related concepts such as sentiment, emotion as well as as higher order constructs such as mood and humor. By its nature, it is highly subjective, with different stimuli leading to different responses in people due to varying personal and cultural artifacts. For example, a specific image or audio clip may evoke different emotions in people depending on their personality. Computational modeling of affective dimensions is an important problem in Artificial Intelligence (AI). It covers a variety of tasks such as sentiment analysis, emotion recognition and opinion mining, which often involve supervised training of models using a large number of labeled data instances. However, training labels are difficult to obtain due to the inherent subjectivity of these constructs. Typical approaches to obtain the training labels include collecting opinions from expert or naive annotators, followed by a suitable aggregation.
In this dissertation, we will present our contributions towards building computational models for noisy annotations of affect, specifically in the aggregation of multidimensional annotations. We propose latent variable models to capture annotator behaviors using additive Gaussian noise and matrix factorization, leading to more accurate estimates of the underlying ground truth. We then apply the joint matrix factorization model to the task of sentence level estimation of psycholinguistic normatives. Finally, we highlight our ongoing efforts in estimating agreement on multidimensional annotations.
Location: Ronald Tutor Hall of Engineering (RTH) - 320
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Defense - Abdullah Alfarrarjeh
Wed, Oct 23, 2019 @ 10:00 AM - 12:00 PM
Thomas Lord Department of Computer Science
University Calendar
PhD Candidate: Abdullah Alfarrarjeh
Committee:
Cyrus Shahabi (chair)
Aiichiro Nakano
C.-C. Jay Kuo
Location: PHE 325
Time: October 23rd, 10 am.
Title: Enabling Spatial-Visual Search for Geospatial Image Databases
Abstract:
Due to continuous advances in camera technologies as well as camera-enabled devices (e.g., CCTV, smartphone, vehicle blackbox, and GoPro), urban streets have been documented by massive amounts of images. Moreover, nowadays, images are typically tagged with spatial metadata due to various sensors (e.g., GPS and digital compass) attached to or embedded in cameras. Such images are known as geo-tagged images. The availability of such geographical context of images enables emerging several image-based smart city applications. Developing such smart city applications requires searching for images, among the massive amounts of collected images, especially to be used for training various machine learning algorithms. Thus, there is an immense need for a data management system for geo-tagged images.
Towards this end, it is paramount to build a data management system that organizes the images in structures that enable searching and retrieving the images efficiently and accurately. On one hand, the data management system should overcome the challenge of lacking an accurate spatial representation of legacy images that were collected without spatial metadata, as well as representing the content of an image accurately using an enriched visual descriptor. On the other hand, the system should also enable efficient storage of images utilizing both their spatial and visual properties and thus their retrieval based on spatial-visual queries. To address these challenges we present a system which includes three integrated modules: a) modeling an image spatially by its scene location using a data-centric approach, b) extending the visual representation of an image with the feature set of multiple similar images located in its vicinity, and c) designing index structures that expedite the evaluation of spatial-visual queries.
Location: 325
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
