Tagged articles
concept drift
11 articles · Page 1 of 1
This article details a systematic process for detecting and managing obsolete experience cards in RAG-based agents, covering drift classification, review metadata, staged deprecation, feedback handling, validation requirements, and in-flight task remediation, with references to Evidently, concept drift research, and PROV-O provenance.
Data DriftLLMOpsPROV-O
0 likes · 22 min read

APO (Autonomous Preference Optimization) converts the drift and conflict among multiple teacher multimodal LLMs into dynamic negative constraints while treating consensus as a positive preference, enabling robust concept alignment and superior diagnostic accuracy on the CXR‑MAX benchmark, as demonstrated by extensive ICML‑2026 experiments.
APOICML 2026concept drift
0 likes · 11 min read

The paper introduces APO, a novel autonomous preference optimization framework that converts concept drift among multiple teacher multimodal LLMs into dynamic negative constraints and treats consensus as a positive preference, achieving robust concept alignment and surpassing strong teachers on a high‑risk medical X‑ray benchmark.
APOCXR-MAXICML 2026
0 likes · 11 min read

The article examines four common yet hidden model evaluation mistakes—confusing attractive metrics with business impact, using static test sets, ignoring statistical significance, and lacking fine‑grained attribution—illustrating each with real‑world cases and offering concrete practices to build a more robust, business‑aligned evaluation pipeline.
A/B testingAI DeploymentMetrics
0 likes · 8 min read

This article reviews a recent paper that proposes a drift‑aware data‑stream system integrating machine‑learning‑based adaptive control into financial data management, introducing a parametric data‑operation module, a gradient‑based bi‑level optimizer, and a curriculum planner to improve model robustness and risk‑adjusted returns in non‑stationary markets.
Data AugmentationQuantitative Financeadaptive data synthesis
0 likes · 18 min read


JD TechApr 1, 2025 · Artificial Intelligence
This article presents a self‑isolation based streaming anomaly detection framework that combines memory‑space indexing to capture pattern anomalies, long‑term memory, and concept drift in time‑series data, and validates the approach with public benchmarks and real‑world risk‑control scenarios.
Time Seriesanomaly detectionconcept drift
0 likes · 24 min read

This article presents a comprehensive analysis of a self‑isolation‑based streaming anomaly detection framework, covering business motivations, existing techniques, technical challenges such as pattern anomalies, long‑term memory and concept drift, the core self‑isolation mechanism, memory‑space architecture, experimental evaluations, and practical risk‑control applications.
Time Seriesanomaly detectionconcept drift
0 likes · 28 min read

This article presents a comprehensive study of streaming‑time‑series anomaly detection, introducing a self‑isolation mechanism combined with a memory space to capture pattern anomalies, handle concept drift, and reduce false alarms, supported by extensive experiments on public datasets and real‑world risk‑control scenarios.
Time Seriesanomaly detectionconcept drift
0 likes · 27 min read

This article presents a comprehensive study of time‑series anomaly detection using a self‑isolation mechanism combined with a memory‑space architecture, addressing pattern anomalies, long‑term memory, and concept drift, and demonstrates its effectiveness through extensive experiments and real‑world risk‑control scenarios.
Time Seriesanomaly detectionconcept drift
0 likes · 28 min read

Even when individual algorithm modules show higher accuracy or revenue, the overall platform can decline due to factors like competitor encroachment, macro‑economic shifts, concept drift, overlapping marginal returns, attribution errors, and coupled A/B experiments, all of which require careful analysis and mitigation.
AB TestingAlgorithmMetrics
0 likes · 7 min read

This article explains a practical, supervised approach to concept‑drift detection by labeling historical data as 0 and new samples as 1, training an LSTM (or LightGBM) classifier, using high‑confidence predictions to flag distribution‑inconsistent samples, and includes code and real‑world examples.
LSTMLightGBMconcept drift
0 likes · 7 min read
