Word embeddings have been shown adept at capturing the semantic and syntactic regularities of the natural language text, as a result of which these representations have found their utility in a wide variety of downstream content analysis tasks. Commonly, these word embedding techniques derive the distributed representation of words based on the local context information. However, such approaches ignore the rich amount of explicit information present in knowledge-bases. This is problematic, as it might lead to poor representation for words with insufficient local context such as domain specific words. Furthermore, the problem becomes pronounced in domain such as bio-medicine where the presence of these domain specific words are relatively high. Towards this end, in this project, we propose a new word embedding based model for biomedical domain that jointly leverages the information from available corpora and domain knowledge in order to generate knowledge-base powered embeddings. Unlike existing approaches, the proposed methodology is simple but adept at capturing the precise knowledge available in domain resources in an accurate way. Experimental results on biomedical concept similarity and relatedness task validates the effectiveness of the proposed approach.

Assessing the complexity of functions computed by a neural network helps us understand how the network will learn and generalize. One natural measure of complexity is how the network distorts length -- if the network takes a unit-length curve as input, what is the length of the resulting curve of outputs? It has been widely believed that this length grows exponentially in network depth. We prove that in fact this is not the case: the expected length distortion does not grow with depth, and indeed shrinks slightly, for ReLU networks with standard random initialization. We also generalize this result by proving upper bounds both for higher moments of the length distortion and for the distortion of higher-dimensional volumes. These theoretical results are corroborated by our experiments, which indicate that length distortion remains modest even after training.

In this paper, we consider fully connected feed-forward deep neural networks where weights and biases are independent and identically distributed according to Gaussian distributions. Extending previous results (Matthews et al., 2018a;b; Yang, 2019) we adopt a function-space perspective, i.e. we look at neural networks as infinite-dimensional random elements on the input space $\mathbb{R}^I$. Under suitable assumptions on the activation function we show that: i) a network defines a continuous Gaussian process on the input space $\mathbb{R}^I$; ii) a network with re-scaled weights converges weakly to a continuous Gaussian process in the large-width limit; iii) the limiting Gaussian process has almost surely locally $\gamma$-H\"older continuous paths, for $0 < \gamma <1$. Our results contribute to recent theoretical studies on the interplay between infinitely wide deep neural networks and Gaussian processes by establishing weak convergence in function-space with respect to a stronger metric.

Generative adversarial neural networks (GANs) [1] [2] [3] [4] are deep learning (DL) models whereby a dataset is used by an agent, called the generator, to sample white noise from a latent space and simulate a data distribution to create new (fake) data that resemble the original data it has been trained on. Another agent, called the discriminator, has to correctly discern between the original data (provided by the external environment for training) and the fake data (produced by the generator). The generator prevails over the discriminator if the latter does not succeed in distinguishing anymore the original from the fake. The discriminator prevails over the generator if the fake data created by the generator is categorized as fake, and the original data is still categorized as original. An illustration that describes a GANs model is shown in Figure 1.

Medical image segmentation models are typically supervised by expert annotations at the pixel-level, which can be expensive to acquire. In this work, we propose a method that combines the high quality of pixel-level expert annotations with the scale of coarse DNN-generated saliency maps for training multi-label semantic segmentation models. We demonstrate the application of our semi-supervised method, which we call CheXseg, on multi-label chest x-ray interpretation. We find that CheXseg improves upon the performance (mIoU) of fully-supervised methods that use only pixel-level expert annotations by 13.4% and weakly-supervised methods that use only DNN-generated saliency maps by 91.2%. Furthermore, we implement a semi-supervised method using knowledge distillation and find that though it is outperformed by CheXseg, it exceeds the performance (mIoU) of the best fully-supervised method by 4.83%. Our best method is able to match radiologist agreement on three out of ten pathologies and reduces the overall performance gap by 71.6% as compared to weakly-supervised methods.

Model-agnostic meta-learning (MAML) has emerged as one of the most successful meta-learning techniques in few-shot learning. It enables us to learn a meta-initialization} of model parameters (that we call meta-model) to rapidly adapt to new tasks using a small amount of labeled training data. Despite the generalization power of the meta-model, it remains elusive that how adversarial robustness can be maintained by MAML in few-shot learning. In addition to generalization, robustness is also desired for a meta-model to defend adversarial examples (attacks). Toward promoting adversarial robustness in MAML, we first study WHEN a robustness-promoting regularization should be incorporated, given the fact that MAML adopts a bi-level (fine-tuning vs. meta-update) learning procedure. We show that robustifying the meta-update stage is sufficient to make robustness adapted to the task-specific fine-tuning stage even if the latter uses a standard training protocol. We also make additional justification on the acquired robustness adaptation by peering into the interpretability of neurons' activation maps. Furthermore, we investigate HOW robust regularization can efficiently be designed in MAML. We propose a general but easily-optimized robustness-regularized meta-learning framework, which allows the use of unlabeled data augmentation, fast adversarial attack generation, and computationally-light fine-tuning. In particular, we for the first time show that the auxiliary contrastive learning task can enhance the adversarial robustness of MAML. Finally, extensive experiments are conducted to demonstrate the effectiveness of our proposed methods in robust few-shot learning.

The graph convolutional network (GCN) is a go-to solution for machine learning on graphs, but its training is notoriously difficult to scale in terms of both the size of the graph and the number of model parameters. These limitations are in stark contrast to the increasing scale (in data size and model size) of experiments in deep learning research. In this work, we propose GIST, a novel distributed approach that enables efficient training of wide (overparameterized) GCNs on large graphs. GIST is a hybrid layer and graph sampling method, which disjointly partitions the global model into several, smaller sub-GCNs that are independently trained across multiple GPUs in parallel. This distributed framework improves model performance and significantly decreases wall-clock training time. GIST seeks to enable large-scale GCN experimentation with the goal of bridging the existing gap in scale between graph machine learning and deep learning.

The understanding of the human language is quantified by identifying intents and entities. Even though classification methods that rely on labeled information are often used for the comprehension of language understanding, it is incredibly time consuming and tedious process to generate high propensity supervised datasets. In this paper, we present the generation of accurate intents for the corresponding Roman Urdu unstructured data and integrate this corpus in RASA NLU module for intent classification. We embed knowledge graph with RASA Framework to maintain the dialog history for semantic based natural language mechanism for chatbot communication. We compare results of our work with existing linguistic systems combined with semantic technologies. Minimum accuracy of intents generation is 64 percent of confidence and in the response generation part minimum accuracy is 82.1 percent and maximum accuracy gain is 96.7 percent. All the scores refers to log precision, recall, and f1 measure for each intents once summarized for all. Furthermore, it creates a confusion matrix represents that which intents are ambiguously recognized by approach.

In this paper, we aim to improve the data efficiency of image captioning. We propose VisualGPT, a data-efficient image captioning model that leverages the linguistic knowledge from a large pretrained language model (LM). A crucial challenge is to balance between the use of visual information in the image and prior linguistic knowledge acquired from pretraining.We designed a novel self-resurrecting encoder-decoder attention mechanism to quickly adapt the pretrained LM as the language decoder on a small amount of in-domain training data. The pro-posed self-resurrecting activation unit produces sparse activations but is not susceptible to zero gradients. When trained on 0.1%, 0.5% and 1% of MSCOCO and Conceptual Captions, the proposed model, VisualGPT, surpasses strong image captioning baselines. VisualGPT outperforms the best baseline model by up to 10.8% CIDEr on MS COCO and up to 5.4% CIDEr on Conceptual Captions.We also perform a series of ablation studies to quantify the utility of each system component. To the best of our knowledge, this is the first work that improves data efficiency of image captioning by utilizing LM pretrained on unimodal data. Our code is available at: https://github.com/Vision-CAIR/VisualGPT.

Standard deep reinforcement learning algorithms use a shared representation for the policy and value function. However, we argue that more information is needed to accurately estimate the value function than to learn the optimal policy. Consequently, the use of a shared representation for the policy and value function can lead to overfitting. To alleviate this problem, we propose two approaches which are combined to create IDAAC: Invariant Decoupled Advantage Actor-Critic. First, IDAAC decouples the optimization of the policy and value function, using separate networks to model them. Second, it introduces an auxiliary loss which encourages the representation to be invariant to task-irrelevant properties of the environment. IDAAC shows good generalization to unseen environments, achieving a new state-of-the-art on the Procgen benchmark and outperforming popular methods on DeepMind Control tasks with distractors. Moreover, IDAAC learns representations, value predictions, and policies that are more robust to aesthetic changes in the observations that do not change the underlying state of the environment.