Amir Rahimi

Samira Kaviani, Amir Rahimi, Richard Hartley

To obtain 3D annotations, we are restricted to controlled environments or synthetic datasets, leading us to 3D datasets with less generalizability to real-world scenarios. To tackle this issue in the context of semi-supervised 3D hand shape and pose estimation, we propose the Pose Alignment network to propagate 3D annotations from labelled frames to nearby unlabelled frames in sparsely annotated videos. We show that incorporating the alignment supervision on pairs of labelled-unlabelled frames allows us to improve the pose estimation accuracy. Besides, we show that the proposed Pose Alignment network can effectively propagate annotations on unseen sparsely labelled videos without fine-tuning.

Amirreza Shaban, Amir Rahimi, Thalaiyasingam Ajanthan et al.

Detecting novel objects from few examples has become an emerging topic in computer vision recently. However, these methods need fully annotated training images to learn new object categories which limits their applicability in real world scenarios such as field robotics. In this work, we propose a probabilistic multiple instance learning approach for few-shot Common Object Localization (COL) and few-shot Weakly Supervised Object Detection (WSOD). In these tasks, only image-level labels, which are much cheaper to acquire, are available. We find that operating on features extracted from the last layer of a pre-trained Faster-RCNN is more effective compared to previous episodic learning based few-shot COL methods. Our model simultaneously learns the distribution of the novel objects and localizes them via expectation-maximization steps. As a probabilistic model, we employ von Mises-Fisher (vMF) distribution which captures the semantic information better than Gaussian distribution when applied to the pre-trained embedding space. When the novel objects are localized, we utilize them to learn a linear appearance model to detect novel classes in new images. Our extensive experiments show that the proposed method, despite being simple, outperforms strong baselines in few-shot COL and WSOD, as well as large-scale WSOD tasks.

Amir Rahimi, Thomas Mensink, Kartik Gupta et al.

Calibration of neural networks is a critical aspect to consider when incorporating machine learning models in real-world decision-making systems where the confidence of decisions are equally important as the decisions themselves. In recent years, there is a surge of research on neural network calibration and the majority of the works can be categorized into post-hoc calibration methods, defined as methods that learn an additional function to calibrate an already trained base network. In this work, we intend to understand the post-hoc calibration methods from a theoretical point of view. Especially, it is known that minimizing Negative Log-Likelihood (NLL) will lead to a calibrated network on the training set if the global optimum is attained (Bishop, 1994). Nevertheless, it is not clear learning an additional function in a post-hoc manner would lead to calibration in the theoretical sense. To this end, we prove that even though the base network ($f$) does not lead to the global optimum of NLL, by adding additional layers ($g$) and minimizing NLL by optimizing the parameters of $g$ one can obtain a calibrated network $g \circ f$. This not only provides a less stringent condition to obtain a calibrated network but also provides a theoretical justification of post-hoc calibration methods. Our experiments on various image classification benchmarks confirm the theory.

Samaneh Aghajanbaglo, Sobhan Moosavi, Maseud Rahgozar et al.

Protein-Protein Interactions (PPIs) perform essential roles in biological functions. Although some experimental techniques have been developed to detect PPIs, they suffer from high false positive and high false negative rates. Consequently, efforts have been devoted during recent years to develop computational approaches to predict the interactions utilizing various sources of information. Therefore, a unique category of prediction approaches has been devised which is based on the protein sequence information. However, finding an appropriate feature encoding to characterize the sequence of proteins is a major challenge in such methods. In presented work, a sequence based method is proposed to predict protein-protein interactions using N-Gram encoding approaches to describe amino acids and a Relaxed Variable Kernel Density Estimator (RVKDE) as a machine learning tool. Moreover, since proteins can rotate in 3D-space, amino acid compositions have been considered with "undirected" property which leads to reduce dimensions of the vector space. The results show that our proposed method achieves the superiority of prediction performance with improving an F-measure of 2.5% on Human Protein Reference Dataset (HPRD).