Neural Information Processing Systems http://nips.cc/

Recently a wide variety of NAS methods have been proposed and achieved considerable success in automatically identifying highly-performing architectures of neural networks for the sake of reducing the reliance on human experts. Existing NAS methods ignore the fact that different input data elements (e.g., image pixels) have different importance (or saliency) in determining the prediction outcome. They treat all data elements as being equally important and therefore lead to suboptimal performance. To address this problem, we propose an end-to-end framework which dynamically detects saliency of input data, reweights data using saliency maps, and searches architectures on saliency-reweighted data. Our framework is based on four-level optimization, which performs four learning stages in a unified way.

Finally, we report empirical evaluation on synthetic and real-world data showing that our proposed SVD-based method does indeed achieve a better cost than other widely-used heurstics and also results in a better classification accuracy when the underlying problem was that of multi-class classification.

We consider massive distributed datasets that consist of elements modeled as key-value pairs and the task of computing statistics or aggregates where the contribution of each key is weighted by a function of its frequency (sum of values of its elements).

Neural Information Processing Systems http://nips.cc/

Abstract: Based on economic theories and integrated with machine learning technology, this study explores the collaborative model of Supply Chain Management - Financial Supply Chain Management (SCM - FSCM), aiming to solve supply chain issues (efficiency l oss, financing constraints, risk transmission) caused by the disconnection of the "three flows" (capital flow, logistics flow, information flow) and further improve overall economic benefits. Firstly, the study combines Transaction Cost Theory and Information Asymmetry Theory, adopts algorithms such as random forests to process multi - dimensional supply chain data, identifies obstacles to the collaboration of the "three flows", and constructs a data - driven three - dimensional (cost - efficiency - risk) a nalysis framework. Secondly, it designs a Financial Supply Chain Management model of "core enterprise credit empowerment + dynamic pledge financing". Based on inventory/order data in Supply Chain Management, it applies Long Short - Term Memory (LSTM) netwo rks to realize demand forecasting, and at the same time uses clustering/regression algorithms to quantify benefit distribution, so as to achieve reasonable allocation of financing costs. In addition, the study also combines Game Theory and reinforcement learning to optimize the supply chain inventory - procurement mechanism (adjusts strategies through scenario simulation to solve problems caused by the "bullwhip effect"); and integrates accounts receivable financing in Financial Supply Chain Management with credit assessment based on eXtreme Gradient Boosting (XGBoost) to realize rapid monetization of inventory.

Weighted sampling is a fundamental tool in data analysis and machine learning pipelines.

Weighted sampling is a fundamental tool in data analysis and machine learning pipelines.