Text Mining Support in Semantic Annotation and Indexing of Multimedia Data

AAAI Conferences

This short paper is describing a demonstrator that is complementing the paper "Towards Cross-Media Feature Extraction" in these proceedings. The demo is exemplifying the use of textual resources, out of which semantic information can be extracted, for supporting the semantic annotation and indexing of associated video material in the soccer domain. Entities and events extracted from textual data are marked-up with semantic classes derived from an ontology modeling the soccer domain. We show further how extracted Audio-Video features by video analysis can be taken into account for additional annotation of specific soccer event types, and how those different types of annotation can be combined.


To Improve Customer Service Robots Enable Humans - InformationWeek

#artificialintelligence

Companies today have a customer service problem, and fixing it is more complicated than flashing an eager smile. Consumer-facing businesses are grappling with how best to meet the fickle expectations of real people in an increasingly automated and digital world. At the center of the issue are automated customer service systems, also called "virtual agents." These agents are software programs designed to help customers answer questions, perform basic tasks, or solve problems without talking to an actual person. We've all used them, and in many cases they work great.


Optimize TSK Fuzzy Systems for Big Data Classification Problems: Bag of Tricks

arXiv.org Machine Learning

Takagi-Sugeno-Kang (TSK) fuzzy systems are flexible and interpretable machine learning models; however, they may not be easily applicable to big data problems, especially when the size and the dimensionality of the data are both large. This paper proposes a mini-batch gradient descent (MBGD) based algorithm to efficiently and effectively train TSK fuzzy systems for big data classification problems. It integrates three novel techniques: 1) uniform regularization (UR), which is a regularization term added to the loss function to make sure the rules have similar average firing levels, and hence better generalization performance; 2) random percentile initialization (RPI), which initializes the membership function parameters efficiently and reliably; and, 3) batch normalization (BN), which extends BN from deep neural networks to TSK fuzzy systems to speedup the convergence and improve generalization. Experiments on nine datasets from various application domains, with varying size and feature dimensionality, demonstrated that each of UR, RPI and BN has its own unique advantages, and integrating all three together can achieve the best classification performance.


Will Artificial Intelligence be a Trump Card for MedTech?

#artificialintelligence

In the era of big data, unless one's portfolio is in tune with the evolving digital trend, making prudent investment choices can prove to be a daunting task. Millennials have started to recognize the increasing need of the emerging automation trend and subsequently robotics, IoT, 3D printing are becoming part of our daily life with the latest buzzword being Artificial Intelligence (AI). While Siri and Ok Google have made life easier, the latest version of Global Positioning System helps track almost anything and everything along with its intelligent route map. Also, not being a pro on social network sites like Facebook, Twitter or snapchat can be seen as primitive. According to investment giant Jim Cramer, AI along with big data will soon let companies to bat a thousand.


21 Scary Things Big Data Knows About You

@machinelearnbot

Of course, Google knows what you've searched for. So do Bing, Yahoo!, and every other search engine. And your ISP knows every website you've ever visited. Google also knows your age and gender -- even if you never told them. They make a pretty comprehensive ads profile of you, including a list of your interests (which you can edit) to decide what kinds of ads to show you.