Tag
The paper proposes a process mining-enabled federated anomaly detection pipeline that uses similarity metrics against a global node to dynamically weight local models during aggregation, addressing non-IID data heterogeneity in industrial IoT. Experiments on manufacturing plant time-series data show 46.52% and 11.99% F1-score improvements over isolated training and FedAvg, plus a 51.24% reduction in false positives.
A paper proposing a multi-layered AI pipeline that converts natural language process descriptions into BPMN diagrams by combining a fine-tuned RoBERTa entity extractor, LLM-based control-flow relation extraction, and ontology population, reporting 90.7% accuracy on relation extraction.
MIT Technology Review 的 Business Lab 播客访谈中,Celonis 首席 AI 科学家 Devavrat Shah 阐述为何企业 AI 需要理解业务运营上下文才能从试验走向规模化落地。
A forward-deployed engineering team audited a real approval workflow at a $5B public software company using process-mining agents on its CRM, uncovering 20 actual steps and 7 loops versus the documented 7 steps, including 61% of requests looping back to the start. The post argues teams should map and fix broken processes before deploying AI.
This guide outlines the playbook behind forward deployed engineers (FDEs) who use AI agents to rebuild legacy workflows inside client companies, from mapping the real process to deploying agents in existing systems of record. It references private equity firms like Thrive Holdings driving large efficiency gains by embedding FDEs across portfolio companies.
The paper derives exact changes in marginal log-likelihood for Dirichlet-smoothed Markov models when adding workflow traces, and evaluates trace selection methods under budget constraints using the BPI Challenge 2012 dataset.
The paper proposes a goal-driven approach to categorize process variants in process mining by using a goal model and Large Language Models (LLMs) to interpret and assign variants to predefined categories, evaluated on public logs.
This paper presents the first comprehensive empirical study on multi-task learning for predictive process monitoring, showing improvements over single-task learning in next-activity prediction and class imbalance mitigation.
Presents SERUM, a multi-pass framework that extracts structured behavioral models of user actions and intents from raw egocentric video using hierarchical VLM annotation, reducing hallucinations and producing interpretable process models without manual annotation.
Trace2Policy extracts human-readable decision rules from expert behavior traces and iteratively refines them via error-driven skill refinement, outperforming pure LLM baselines on compliance-sensitive tasks in logistics.
This paper introduces a totally unimodular linear programming reformulation for alignment-based conformance checking, which complements A* search by providing speedups for long traces with deviations. The approach achieves 38.6% average runtime savings with 96% selection accuracy.
Flowscope is a Y Combinator-backed AI-native consulting firm that deploys agents to map, redesign, and automate business processes within days by integrating directly into existing enterprise systems.
本文深度解析了 Ramp Labs 研发的 AI 代理电子表格工具 RAMP Sheets,涵盖其从流程挖掘起步的演进路径、基于 Excel 公式优先的透明化智能体架构设计,以及在财务自动化场景中的实际应用。