Recognition: unknown
Estimate and Replace: A Novel Approach to Integrating Deep Neural Networks with Existing Applications
read the original abstract
Existing applications include a huge amount of knowledge that is out of reach for deep neural networks. This paper presents a novel approach for integrating calls to existing applications into deep learning architectures. Using this approach, we estimate each application's functionality with an estimator, which is implemented as a deep neural network (DNN). The estimator is then embedded into a base network that we direct into complying with the application's interface during an end-to-end optimization process. At inference time, we replace each estimator with its existing application counterpart and let the base network solve the task by interacting with the existing application. Using this 'Estimate and Replace' method, we were able to train a DNN end-to-end with less data and outperformed a matching DNN that did not interact with the external application.
This paper has not been read by Pith yet.
Forward citations
Cited by 5 Pith papers
-
Muon is Not That Special: Random or Inverted Spectra Work Just as Well
Muon succeeds by guaranteeing local step-size optimality rather than by tracking any ideal global geometry, as random-spectrum and quasi-norm variants match its performance on language models.
-
Letting the neural code speak: Automated characterization of monkey visual neurons through human language
Natural-language descriptions generated and verified through generative models and digital twins capture the selectivity of most neurons in macaque V1 and V4.
-
Liberata -- Graph Scientometrics for a Share Based System of Academic Publishing
Liberata introduces a graph-based system using continuous contribution shares and weighted citations to derive metrics for impact, risk, collaboration, and quality control in academic publishing.
-
Detecting Language Model Attacks with Perplexity
Jailbreak prompts with adversarial suffixes have high GPT-2 perplexity, and a LightGBM model on perplexity and length detects most attacks.
-
Student Classroom Behavior Recognition Based on Improved YOLOv8s
An updated YOLOv8s detector improves student behavior recognition in crowded classrooms by 1.8-2.1% mAP through added feature modules and a reweighted loss.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.