Abstract

Optical coherence tomographic (OCT) images provide valuable information for understanding the changes occurring in the retina due to glaucoma, specifically, related to the retinal nerve fiber layer and the optic nerve head. In this paper, we propose a deep learning approach using Capsule network for glaucoma classification, which directly operates on 3D OCT volumes. The network is trained only on labelled volumes and does not attempt any region/structure segmentation. The proposed network was assessed on 50 volumes and found to achieve 0.97 for the area under the ROC curve (AUC). This is considerably higher than the existing approaches which are majorly based on machine learning or rely on segmentation of the required structures from OCT. Our network also outperforms 3D convolutional neural networks despite the fewer network parameters and fewer epochs needed for training.

Abstract

Localizing natural language phrases in images is a challenging problem that requires joint understanding of both the textual and visual modalities. In the unsupervised setting, lack of supervisory signals exacerbate this difficulty. In this paper, we propose a novel framework for unsupervised visual grounding which uses concept learning as a proxy task to obtain self-supervision. The intuition behind this idea is to encourage the model to localize to regions which can explain some semantic property in the data, in our case, the property being the presence of a concept in a set of images. We present thorough quantitative and qualitative experiments to demonstrate the efficacy of our approach and show a 5.6% improvement over the current state of the art on Visual Genome dataset, a 5.8% improvement on the ReferItGame dataset and comparable to state-of-art performance on the Flickr30k dataset.

Abstract

We consider a setting where public resources are to be allocated among competing and strategic agents so as to maximize social welfare (the objects should be allocated to those who value them the most). This is called allocative efficiency (AE). We need the agents to report their valuations for obtaining these resources, truthfully referred to as dominant strategy incentive compatibility (DSIC). We use auction-based mechanisms to achieve AE and DSIC yet budget balance cannot be ensured, due to Green-Laffont Impossibility Theorem. That is, the net transfer of money cannot be zero. This problem has been addressed by designing a redistribution mechanism so as to ensure a minimum surplus of money as well as AE and DSIC. The objective could be to minimize surplus in expectation or in the worst case and these objects could be homogeneous or heterogeneous. Designing redistribution mechanisms which perform well in expectation becomes analytically challenging for heterogeneous settings. In this paper, we take a completely different, data-driven approach. We train a neural network to determine an optimal redistribution mechanism based on given settings with both the objectives, optimal in expectation and optimal in the worst case. We also propose a loss function to train a neural network to optimize worst case. We design neural networks with the underlying rebate functions being linear as well as nonlinear in terms of bids of the agents. Our networks' performances are same as the theoretical guarantees for the cases where it has been solved. We observe that a neural network based redistribution mechanism for homogeneous …

Abstract

Monocular head pose estimation requires learning a model that computes the intrinsic Euler angles for pose (yaw, pitch, roll) from an input image of human face. Annotating ground truth head pose angles for images in the wild is difficult and requires ad-hoc fitting procedures (which provides only coarse and approximate annotations). This highlights the need for approaches which can train on data captured in controlled environment and generalize on the images in the wild (with varying appearance and illumination of the face). Most present day deep learning approaches which learn a regression function directly on the input images fail to do so. To this end, we propose to use a higher level representation to regress the head pose while using deep learning architectures. More specifically, we use the uncertainty maps in the form of 2D soft localization heatmap images over five facial keypoints, namely left ear, right ear, left eye, right eye and nose, and pass them through an convolutional neural network to regress the head-pose. We show head pose estimation results on two challenging benchmarks BIWI and AFLW and our approach surpasses the state of the art on both the datasets. Index Terms— Image analysis, Pose estimation

Abstract

We describe the design of an autonomous electricity broker agent, VidyutVanika, the runner-up of the 2018 PowerTAC competition. The agent uses techniques from reinforcement learning, dynamic programming and other areas of machine learning to seek appropriate actions in tariff and wholesale market of the PowerTAC simulation environment. The novelty of our agent lies in defining the reward functions of suitably defined Markov decision processes (MDPs), solving these MDPs, and applying these solutions to real actions in the market. In addition, VidyutVanika uses a neural network to predict the energy consumption of various customers using weather data. The usage forecasts, so obtained, are used to place orders in day-ahead wholesale market. These forecasts also helps in reducing the balancing costs incurred by the broker.

Abstract

Methods and systems are disclosed for providing cloud services to multiple customers in a cloud. One embodiment includes receiving a number of requests for the cloud services from the multiple customers simultaneously or substantially simultaneously; prioritizing the requests based on a probability distribution of actually deploying a service, a budget of the customers, and an expected demand of the requested service based on the probability distribution; generating a number of cloud configurations along with a number of Service Level Agreements (SLAs) for the customers based on prioritization of the requests, a class & past behavior of the customers, and a current demand of the cloud services, the SLAs of the customers include differentiated price offering; recommending the cloud configurations and the SLAs to the customers; allowing the customers to negotiate terms of the SLAs; and providing the cloud …

Abstract

Abstract— We propose a novel pipeline that blends encodings from natural language and 3D semantic maps obtained from visual imagery to generate local trajectories that are executed by a low-level controller. The pipeline precludes the need for a prior registered map through a local waypoint generator neural network. The waypoint generator network (WGN) maps semantics and natural language encodings (NLE) to local waypoints. A local planner then generates a trajectory from the ego location of the vehicle (an outdoor car in this case) to these locally generated waypoints while a low-level controller executes these plans faithfully. The efficacy of the pipeline is verified in the CARLA simulator environment as well as on local semantic maps built from real-world KITTI dataset. In both these environments (simulated and real-world) we show the ability of the WGN to generate waypoints accurately by mapping NLE of varying sequence lengths and levels of complexity. We compare with baseline approaches and show significant performance gain over them. And finally, we show real implementations on our electric car verifying that the pipeline lends itself to practical and tangible realizations in uncontrolled outdoor settings. In loop execution of the proposed pipeline that involves repetitive invocations of the network is critical for any such language-based navigation framework. This effort successfully accomplishes this thereby bypassing the need for prior metric maps or strategies for metric level localization during traversal.

Abstract

In many practical applications such as crowd-sourcing and online advertisement, use of mechanism design (auction-based mechanisms) depends upon inherent stochastic parameters which are unknown. These parameters are learnt using multi-armed bandit (MAB) algorithms. The mechanisms which incorporate MAB are referred to as Multi-Armed-Bandit Mechanisms. While most of the MAB mechanisms focus on frequentist approaches like upper confidence bound algorithms, recent work has shown that using Bayesian approaches like Thompson sampling results in mechanisms with better regret bounds; although lower regret is obtained at the cost of the mechanism ending up with a weaker game theoretic property ie Within-Period Dominant Strategy Incentive Compatibility (WP-DSIC). The existing payment rules used in the Thompson sampling based mechanisms may cause negative utility to the auctioneer. In addition, if we wish to minimize the cost to the auctioneer, it is very challenging to design payment rules that satisfy WP-DSIC while learning through Thompson sampling. In our work, we propose a data-driven approach for designing MAB-mechanisms. Specifically, we use neural networks for designing the payment rule which is WP-DSIC, while the allocation rule is modeled using Thompson sampling. Our results, in the setting of crowd-sourcing for recruiting quality workers, indicate that the learned payment rule guarantees better cost while maximizing the social welfare and also ensuring reduced variance in the utilities to the agents.

Abstract

We focus on the aggregation of citizen preferences for public projects through civic crowdfunding. Existing civic crowdfunding mechanisms consider only agents with positive valuation towards the public project. Moreover, these mechanisms assume that each agent has a symmetric belief about the project getting provisioned. As public projects aim to cater to the majority, they should be provisioned only if the majority prefers it. To incorporate negative valuations, we propose a methodology to convert existing civic crowdfunding mechanisms for positive preferences to cater to markets having both types of agents. Specifically, we adapt existing PPR and PPS mechanisms to design PPRN and PPSN, that incentivize agents to contribute towards or against the project’s provision: based on their preference. Besides, to address asymmetric beliefs, we propose a novel reward scheme, Belief Based Reward (BBR) based on Robust Bayesian Truth Serum (RBTS) mechanism. BBR rewards agents based on their belief towards the project’s provision. Using this reward scheme, we introduce a general mechanism for civic crowdfunding which allows for agents having asymmetric beliefs towards the project getting provisioned and incentivizes them to contribute towards the project’s provision. We illustrate the general mechanism by designing two novel mechanisms, namely PPRx and PPSx, adapting PPR and PPS respectively, and prove that in both the mechanisms the project is provisioned at equilibrium

Abstract

Auctions are mechanisms which facilitate the buying and selling of goods/items amongst a group of agents. In general, a combinatorial auction, where the agents can bid for combination (s) of items, yields a higher revenue than selling the goods/items individually.