In recent years, significant progress has been made in solving challenging problems across various domains using deep reinforcement learning (RL). Reproducing existing work and accurately judging the improvements offered by novel methods is vital to sustaining this progress. Unfortunately, reproducing results for state-of-the-art deep RL methods is seldom straightforward. In particular, non-determinism in standard benchmark environments, combined with variance intrinsic to the methods, can make reported results tough to interpret. Without significance metrics and tighter standardization of experimental reporting, it is difficult to determine whether improvements over the prior state-of-the-art are meaningful. In this paper, we investigate challenges posed by reproducibility, proper experimental techniques, and reporting procedures. We illustrate the variability in reported metrics and results when comparing against common baselines and suggest guidelines to make future results in deep RL more reproducible. We aim to spur discussion about how to ensure continued progress in the field by minimizing wasted effort stemming from results that are non-reproducible and easily misinterpreted.

In recent years, significant progress has been made in solving challenging problems across various domains using deep reinforcement learning (RL). Reproducing existing work and accurately judging the improvements offered by novel methods is vital to sustaining this progress. Unfortunately, reproducing results for state-of-the-art deep RL methods is seldom straightforward. In particular, non-determinism in standard benchmark environments, combined with variance intrinsic to the methods, can make reported results tough to interpret. Without significance metrics and tighter standardization of experimental reporting, it is difficult to determine whether improvements over the prior state-of-the-art are meaningful. In this paper, we investigate challenges posed by reproducibility, proper experimental techniques, and reporting procedures. We illustrate the variability in reported metrics and results when comparing against common baselines and suggest guidelines to make future results in deep RL more reproducible. We aim to spur discussion about how to ensure continued progress in the field by minimizing wasted effort stemming from results that are non-reproducible and easily misinterpreted.

No abstract

Fine-grained categorization of object classes is receiving increased attention, since it promises to automate classification tasks that are difficult even for humans, such as the distinction between different animal species. In this paper, we consider fine-grained categorization for a different reason: following the intuition that fine-grained categories encode metric information, we aim to generate metric constraints from fine-grained category predictions, for the benefit of 3D scene-understanding. To that end, we propose two novel methods for fine-grained classification, both based on part information, as well as a new fine-grained category data set of car types. We demonstrate superior performance of our methods to state-of-the-art classifiers, and show first promising results for estimating the depth of objects from fine-grained category predictions from a monocular camera.

No abstract

This paper presents a spatial-semantic modeling system featuring automated learning of object-place relations from an online annotated database, and the application of these relations to a variety of real-world tasks. The system is able to label novel scenes with place information, as we demonstrate on test scenes drawn from the same source as our training set. We have designed our system for future enhancement of a robot platform that performs state-of-the-art object recognition and creates object maps of realistic environments. In this context, we demonstrate the use of spatial-semantic information to perform clustering and place labeling of object maps obtained from real homes. This place information is fed back into the robot system to inform an object search planner about likely locations of a query object. As a whole, this system represents a new level in spatial reasoning and semantic understanding for a physical platform.

No abstract

This paper presents a technique to locate objects in 3D that adapts visual appearance models using explicit visibility analysis. We formulate a Bayesian model for 3D object likelihood based on visual appearance, 3D geometry such as that available from RGB-depth sensors, and structure-from-motion. Learned visual appearance templates for portions of an object allow for strong discrimination even under occlusion. We describe an efficient inference procedure based on data-driven sampling with geometric refinement. Our 3D object detection technique is demonstrated on the publicly available robot-collected UBC Visual Robot Survey dataset, as well as with data from the Microsoft Kinect. Results show that our method improves robustness to occlusion when compared to a state-of-the-art visual category detector.