REVIEW 23 cited by
Estimate and Replace: A Novel Approach to Integrating Deep Neural Networks with Existing Applications
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
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.
Forward citations
Cited by 23 Pith papers
-
SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm
SARLO-80 is a new public dataset of 119566 complex SAR-optical-text triplets standardized to 80cm slant-range resolution from 257 locations across 72 countries.
-
Global Explanations for Multivariate Time Series Forecasting Models via $K$-Order Markov Approximations
KARMA constructs minimal-K Markov transition kernels as surrogates to deliver global explanations for multivariate time series forecasting models and recovers known causal structure on synthetic data.
-
SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm
SARLO-80 provides 119,566 complex SAR-optical-text triplets at 80 cm slant-range resolution with fixed splits and preprocessing code.
-
Silicate cosmic dust grain collisions in the interstellar medium: A molecular dynamics study
MD simulations of 5-50 Å silicate grains find shattering thresholds of ~6 km/s for both SiO2 and astrodust compositions, twice the canonical 2.7 km/s value, with shattered size distributions inconsistent with prior po...
-
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.
-
Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination
Graph-PRefLexOR fine-tunes graph-native models with GRPO to organize reasoning into phases, yielding 40-65% gains in traceable hypothesis generation and 2-3x semantic diversity on 100 materials science questions.
-
Data and Evaluation Closed-Loop for Model Capability Enhancement
Proposes capability slices with dual taxonomies and mapping rules to form a closed loop converting benchmark failures into targeted data interventions, validated via two opposing case studies on BBH and math reasoning.
-
A Prototypical Signature Approach for Writer-Independent Offline Signature Verification
Prototypical signatures enable generation of diverse negative samples for writer-independent offline signature verification, improving skilled forgery detection and allowing scalable linear SVM alternatives to RBF models.
-
Provably Efficient Personalized Multi-Objective Bandits with Proactive Conversational Queries
MO-PQUCB hybrid algorithm integrates proactive conversational queries with bandit feedback via shift-invariant regularization to achieve improved regret bounds in personalized multi-objective bandits.
-
What Does Debiasing Really Remove? A Geometric Study of PCA-Based Gender Debiasing in Word Embeddings
Geometric study shows PCA debiasing removes direct gender bias from the first PC but leaves WEAT-measured associative bias distributed and degrades embedding geometry with no universal optimum.
-
Learning Control-Affine Reduced-Order Models via Autoencoders
Framework learns control-affine reduced-order models via jointly trained autoencoders on high-dimensional data, extended to sequence-based models and assessed on numerical examples for prediction and feedback linearization.
-
Silicate cosmic dust grain collisions in the interstellar medium: A molecular dynamics study
MD simulations yield silicate grain shattering thresholds of ~6 km/s and post-collision size distributions inconsistent with power-law predictions from Jones et al. (1996) and Hirashita & Kobayashi (2013).
-
When Support Escalates Distress: Regulation and Escalation in LLM Responses to Venting and Advice-Seeking
LLM responses mirror venting with higher regulation and escalation; therapist personas lower escalation while preserving regulation, and lay raters miss escalation.
-
Stationary subspace analysis for spatial data
Introduces spSSA extending SSA to spatial data via three generalized eigenvalue procedures and a data augmentation method to estimate nonstationary subspace dimension.
-
Symplectic Neural Operators for Learning Infinite Dimensional Hamiltonian Systems
Symplectic Neural Operators preserve symplectic structure for learning infinite-dimensional Hamiltonian PDEs and deliver improved long-term energy stability in theory and experiments.
-
Letting the neural code speak: Automated characterization of monkey visual neurons through human language
Natural language descriptions generated via a closed-loop pipeline with digital twins capture the selectivity of most neurons in macaque V1 and V4, with synthesized images driving 96% of V4 neurons into the top or bot...
-
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.
-
Sequential Probability Ratio Test using Z-Statistics (SPRT-z): A Practical Approach for Online Experimentation
SPRT-z adapts SPRT for online experiments via z-approximation, SFHC Monte Carlo horizon setting, and Brownian median unbiased estimators, with simulations claiming error control, smaller samples than fixed-horizon tes...
-
How Optimality Structures Sparse Dictionaries: A Theory for Understanding SAE Representations
Derives optimality constraints for nonnegative joint dictionary learning that explain observed SAE behaviors such as feature splitting, absorption, and dense antipodal features.
-
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.
-
A critical comparison of handling zeros in high-dimensional compositional count data
A review consolidating zero-handling methods for compositional count data, critiquing log-ratio assumptions, and comparing imputation strategies adapted to discrete zero-inflated counts.
-
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.