An official website of the United States government
Recent advances in computer vision for space exploration have handled prediction uncertainties well by approximating multimodal output distribution rather than averaging the distribution. While those advanced multimodal deep learning models could enhance the scientific and engineering value of autonomous systems by making the optimal decisions in uncertain environments, sequential learning of those approximated information has depended on unimodal or bimodal probability distribution. In a sequence of information learning and transfer decisions, the traditional reinforcement learning cannot accommodate the noise in the data that could be useful for gaining information from other locations, thus cannot handle multimodal and multivariate gains in their transition function. Still, there is a lack of interest in learning and transferring multimodal space information effectively to maximally remove the uncertainty. In this study, a new information theory overcomes the traditional entropy approach by actively sensing and learning information in a sequence. Particularly, the autonomous navigation of a team of heterogeneous unmanned ground and aerial vehicle systems in Mars outperforms benchmarks through indirect learning.
more »
« less
AI and Data-Driven In-situ Sensing for Space Digital Twin
Recent advances in computer vision for space exploration have handled prediction uncertainties well by approximating multimodal output distribution rather than averaging the distribution. While those advanced multimodal deep learning models could enhance the scientific and engineering value of autonomous systems by making the optimal decisions in uncertain environments, sequential learning of those approximated information has depended on unimodal or bimodal probability distribution. In a sequence of information learning and transfer decisions, the traditional reinforcement learning cannot accommodate the noise in the data that could be useful for gaining information from other locations, thus cannot handle multimodal and multivariate gains in their transition function. Still, there is a lack of interest in learning and transferring multimodal space information effectively to maximally remove the uncertainty. In this study, a new information theory overcomes the traditional entropy approach by actively sensing and learning information in a sequence. Particularly, the autonomous navigation of a team of heterogeneous unmanned ground and aerial vehicle systems in Mars outperforms benchmarks through indirect learning.
more »
« less
- Award ID(s):
- 1910397
- PAR ID:
- 10465568
- Date Published:
- Journal Name:
- 2023 IEEE Space Computing Conference (SCC)
- Page Range / eLocation ID:
- 11 to 11
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Estimating multimodal distributions of travel times from real-world data is critical for understanding and managing congestion. Mixture models can estimate the overall distribution when distinct peaks exist in the probability density function, but no transfer of mixture information under epistemic uncertainty across different spatiotemporal scales has been considered for capturing unobserved heterogeneity. In this paper, a physics-informed and -regularized prediction model is developed that shares observations across similarly distributed network segments across time and space. By grouping similar mixture models, the model uses a particular sample distribution at distant non-contiguous unexplored locations and improves TT prediction. Compared to traditional prediction without those updates, the proposed model's 19% of performance show the benefit of indirect learning. Different from traditional travel time prediction tools, the developed model can be used by traffic and planning agencies in knowing how far back in history and what sample size of historic data would be useful for current prediction.more » « less
-
Kim, Been (Ed.)Integrating and processing information from various sources or modalities are critical for obtaining a comprehensive and accurate perception of the real world in autonomous systems and cyber-physical systems. Drawing inspiration from neuroscience, we develop the Information-Theoretic Hierarchical Perception (ITHP) model, which utilizes the concept of information bottleneck. Different from most traditional fusion models that incorporate all modalities identically in neural networks, our model designates a prime modality and regards the remaining modalities as detectors in the information pathway, serving to distill the flow of information. Our proposed perception model focuses on constructing an effective and compact information flow by achieving a balance between the minimization of mutual information between the latent state and the input modal state, and the maximization of mutual information between the latent states and the remaining modal states. This approach leads to compact latent state representations that retain relevant information while minimizing redundancy, thereby substantially enhancing the performance of multimodal representation learning. Experimental evaluations on the MUStARD, CMU-MOSI, and CMU-MOSEI datasets demonstrate that our model consistently distills crucial information in multimodal learning scenarios, outperforming state-of-the-art benchmarks. Remarkably, on the CMU-MOSI dataset, ITHP surpasses human-level performance in the multimodal sentiment binary classification task across all evaluation metrics (i.e., Binary Accuracy, F1 Score, Mean Absolute Error, and Pearson Correlation).more » « less
-
Predictive routing is effective in knowledge transfer. However, it ignores information gained from probability distributions with more than one peak. We introduce traffic multimodal information learning, a new class of transportation decision-making models that can learn and transfer online information from multiple simultaneous observations of a probability distribution with multiple peaks or multiple outcome variables from one time stage to the next. Multimodal learning improves the scientific and engineering value of autonomous vehicles by determining the best routes based on the intended level of exploration, risk, and limits.more » « less
-
To close the gap between current distribution operations and today’s customer expectations, firms need to think differently about how resources are acquired, managed and allocated to fulfill customer requests. Rather than optimize planned resource capacity acquired through ownership or long- term partnerships, this work focuses on a specific supply-side innovation – on-demand distribution platforms. On-demand distribution systems move, store, and fulfill goods by matching autonomous suppliers' resources (warehouse space, fulfillment capacity, truck space, delivery services) to requests on-demand. On-demand warehousing systems can provide resource elasticity by allowing capacity decisions to be made at a finer granularity (at the pallet-level) and commitment (monthly versus yearly), than construct or lease options. However, such systems are inherently more complex than traditional systems, as well as have varying costs and operational structures (e.g., higher variable costs, but little or no fixed costs). New decision- supporting models are needed to capture these trade-offs.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/SCC57168.2023.00010
