Articles — Recent

Abdallah Meddah's CV. Created with the AltaCV template.

Aadhithya Dinesh's Resume

Ali Özen's Resume. Created using the Deedy CV template.

Paper presented at ICCV 2019. This paper targets the task with discrete and periodic class labels (e.g., pose/orientation estimation) in the context of deep learning. The commonly used cross-entropy or regression loss is not well matched to this problem as they ignore the periodic nature of the labels and the class similarity, or assume labels are continuous value. We propose to incorporate inter-class correlations in a Wasserstein training framework by pre-defining (i.e., using arc length of a circle) or adaptively learning the ground metric. We extend the ground metric as a linear, convex or concave increasing function w.r.t. arc length from an optimization perspective. We also propose to construct the conservative target labels which model the inlier and outlier noises using a wrapped unimodal-uniform mixture distribution. Unlike the one-hot setting, the conservative label makes the computation of Wasserstein distance more challenging. We systematically conclude the practical closed-form solution of Wasserstein distance for pose data with either one-hot or conservative target label. We evaluate our method on head, body, vehicle and 3D object pose benchmarks with exhaustive ablation studies. The Wasserstein loss obtaining superior performance over the current methods, especially using convex mapping function for ground metric, conservative label, and closed-form solution.

Group Isomorphisms

A ranking system for tech stocks and bank stocks of S&P 500 companies

Shaila Ang's CV Created with the CV for Freshers template.

In the last few years the resolution of NLP tasks with architectures composed of neural models has taken vogue. There are many advantages to using these approaches especially because there is no need to do features engineering. In this paper, we make a survey of a Deep Learning architecture that propose a resolutive approach to some classical tasks of the NLP. The Deep Learning architecture is based on a cutting-edge model that exploits both word-level and character-level representations through the combination of bidirectional LSTM, CNN and CRF. This architecture has provided cutting-edge performance in several sequential labeling activities for the English language. The architecture that will be treated uses the same approach for the Italian language. The same guideline is extended to perform a multi-task learning involving PoS labeling and sentiment analysis. The results show that the system performs well and achieves good results in all activities. In some cases it exceeds the best systems previously developed for Italian.

A template for muletter, a document class for the typesetting of letters at the Masaryk Univer­sity (Brno, Czech Repub­lic).