Showing 1–2 of 2 results for author: Grande, G
-
Realistic PointGoal Navigation via Auxiliary Losses and Information Bottleneck
Authors:
Guillermo Grande,
Dhruv Batra,
Erik Wijmans
Abstract:
We propose a novel architecture and training paradigm for training realistic PointGoal Navigation -- navigating to a target coordinate in an unseen environment under actuation and sensor noise without access to ground-truth localization. Specifically, we find that the primary challenge under this setting is learning localization -- when stripped of idealized localization, agents fail to stop preci…
▽ More
We propose a novel architecture and training paradigm for training realistic PointGoal Navigation -- navigating to a target coordinate in an unseen environment under actuation and sensor noise without access to ground-truth localization. Specifically, we find that the primary challenge under this setting is learning localization -- when stripped of idealized localization, agents fail to stop precisely at the goal despite reliably making progress towards it. To address this we introduce a set of auxiliary losses to help the agent learn localization. Further, we explore the idea of treating the precise location of the agent as privileged information -- it is unavailable during test time, however, it is available during training time in simulation. We grant the agent restricted access to ground-truth localization readings during training via an information bottleneck. Under this setting, the agent incurs a penalty for using this privileged information, encouraging the agent to only leverage this information when it is crucial to learning. This enables the agent to first learn navigation and then learn localization instead of conflating these two objectives in training. We evaluate our proposed method both in a semi-idealized (noiseless simulation without Compass+GPS) and realistic (addition of noisy simulation) settings. Specifically, our method outperforms existing baselines on the semi-idealized setting by 18\%/21\% SPL/Success and by 15\%/20\% SPL in the realistic setting. Our improved Success and SPL metrics indicate our agent's improved ability to accurately self-localize while maintaining a strong navigation policy. Our implementation can be found at https://github.com/NicoGrande/habitat-pointnav-via-ib.
△ Less
Submitted 17 September, 2021;
originally announced September 2021.
-
Pressure independence of granular flow through an aperture
Authors:
Maria Alejandra Aguirre,
Juan G. Grande,
Adriana Calvo,
Luis A. Pugnaloni,
Jean-Christophe Géminard
Abstract:
We experimentally demonstrate that the flow rate of granular material through an aperture is controlled by the exit velocity imposed to the particles and not by the pressure at the base, contrary to what is often assumed in previous works. This result is achieved by studying the discharge process of a dense packing of monosized disks through an orifice. The flow is driven by a conveyor belt. This…
▽ More
We experimentally demonstrate that the flow rate of granular material through an aperture is controlled by the exit velocity imposed to the particles and not by the pressure at the base, contrary to what is often assumed in previous works. This result is achieved by studying the discharge process of a dense packing of monosized disks through an orifice. The flow is driven by a conveyor belt. This two-dimensional horizontal setup allows to uncouple pressure and velocity and, therefore, to independently control the velocity at which the disks escape the horizontal silo and the pressure in the vicinity of the aperture. The flow rate is found to be directly proportional to the belt velocity, independent of the amount of disks in the container and, thus, independent of the pressure in the outlet region. In addition, this specific experimental configuration makes it possible to get information on the system dynamics from a single image of the disks that rest on the conveyor belt after the discharge.
△ Less
Submitted 17 May, 2010;
originally announced May 2010.
