PinSieve: Production Selective VLM Serving and a Governed Memory Flywheel for Enterprise Content-Quality Triage
cs.LG
Submitted: 2026-08-25
Updated: 2026-08-25
Comments: Accepted at the KDD 2026 workshop "Enterprise AI Agents: From Prototypes to Production."
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- COEF-VQ: Cost-Efficient Video Quality Understanding through a Cascaded Multimodal LLM Framework
- Distilling the Knowledge in a Neural Network
- VisualBERT: A Simple and Performant Baseline for Vision and Language
- ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks
- DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
- Scalable Data Classification for Security and Privacy
- Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks