Universitat Politècnica de Catalunya

deep learning computer vision recurrent neural networks convolutional neural networks generative adversarial networks video processing object detection visual saliency unsupervised learning natural language processing video retrieval neural networks video generative models retrieval multimedia medical imaging visual question answering artificial intelligence image classification audio attention models multimodal deep learning gan video summarization self-supervised learning machine learning imagenet semantic segmentation image processing reinforcement learning perceptron instance segmentation instance retrieval neural machine translation visualization image segmentation adversarial training figure-ground segmentation eeg transfer learning speech synthesis backpropagation object tracking clustering speech recognition lifelogging deep belief network architectures variational autoencoder dqn speech sign language optimization affective computing object candidates upc interpretability wavenet eye fixation object segmentation vision lifelong learning image captioning policy vae human computing crowdsourcing saliency mediaeval search broadcasters archive user interaction wearable cameras spatial transformer video analysis egocentric vision tensorflow domain adaptation face recognition restricted boltzmann machine training nlp video object segmentation image retrieval cross-modal learning keras indexing inception backward propagation speaker identification classification nmt activity locatization optical flow dataset ranking ai hype social 3d convolution event detection person retrieval barcelona barcelonatech face detection annotations 3d reconstruction gnn rbm graph neural networks autoencoder reranking video annotation word embeddings video indexing google web toolkit hpc resnet hate speech gradient descent skip rnn wearables diversity surf event recognition visual scanpath moderation sentiment prediction incremental learning loss function autoregressive models lstm teaching search engine cnn multilayer perceptron pixelcnn computer segmentation diffusion models hierarchical partitions memes deep neural networks brain coding object 3d streaming media iphone http mysql linux ios crowdmm acmmm social event photo clustering instance search python lifeblogging relevance feedback rvos ccma davis visual dialog visual descriptors pattern recognition algorithm email brain-computer interfaces image electroencephalography nearest neighbor bundling interest points rapid serial visual presentation digital images images dementia mattnet web mutual reinforcement algorithm mediavela q-learning workshop pixable time series seq2seq representation learning explainable ai columbia regions phd thesis xai hyperlinking bag of features minecraft visual grounding trecvid signal bci nist interface twitter television interactive microblogging tv labeling game genai generative learning endoscopy etsetb telecom mobilitat erasmus javascript web toolkit wt c++ mild cognitive impairment html web service ai for social good perpcetron web interface semantic shots image edge detection image representation professional documentalists video signal processing semimanual solution moco automatic keyframe selection companies autonomous driving panoptic segmentation broadcasting single representative keyframe algorithm design and analysis multimedia communication ethics language action classification action detection motion estimation cbir deep captioning remote sensing activity recognition lipreading self-learning soundnet sonorization dynamic computation iclr2018 local geometry language model visual localization googlenet skip connections deep q-network network in network densenet nin skip thought video segmentation pixelrnn dbn adaptive computation time methodology error function rework supercomputing gru lip reading softmax logistic regression linear regression active learning interestingness higher education rgbd multiview 3d images depth joint embeddings action recognition software t-sne epoch batch visual reasoning astronomy space policy gradients incrmental learning theano software development catastrophic forgetting normalizing flows caption natural language memorability neural architectures vgg eye tracker automl resnext attributes college outreach aprenentatge automàtic inteligència artificial robots convnet social networks open source alzheimer narrative computational graph chain rule referring expressions word2vec biometrics data partition set learning vanishing gradient 3d analysis point clouds relu mlp catalunya catalonia cgan colorization cross-entropy batch normalization mini-batch location retrieval ealy stopping coclustering adam sgd search engine optimization cloud computing markov decision proccess data augmentation value value function computing