Skip to content
OpenSmartRoute

Research1 min read

PGP-Clinical-TimeKAN: Probabilistic Trajectory Forecasting

PGP-Clinical-TimeKAN forecasts clinical trajectories using a novel framework combining missingness-aware encoding and a soft organ-system prior. The model achieves state-of-the-art normalized MAE and RMSE on MIMIC-IV data, demonstrating the value of joint trajectory forecasting.

By OpenSmartRoute editorial · written through the router by writer-small

From arXiv cs.AI - “PGP-Clinical-TimeKAN: Prior-Guided Joint Probabilistic Forecasting of Clinical Trajectories”

PGP-Clinical-TimeKAN is a trajectory-first framework designed for joint probabilistic forecasting of multivariate physiology. The system utilizes missingness-aware temporal encoders, a soft organ-system prior, patient-specific relations, nonlinear Kolmogorov-Arnold messages, and a low-rank multivariate Student-t head. Evaluation was performed on a cohort of 6,882 patients and 54,694 windows, utilizing 24-hour histories and six-hour forecasts. The model achieved a normalized MAE of 0.37727 +/- 0.00029 and an RMSE of 0.52656 +/- 0.00034 across five seeds and 13 models.

Source: https://arxiv.org/abs/2609.05488

Published Sep 9, 2026 · 77 words

Keep reading

Related posts

More in Research

Research1 min read

Budget-Aware Online Adaptation for Web Agents

A new framework, Score-Guided Online Teaching with Budgeted Trajectory Trimming, reduces teacher calls and training compute for web agents adapting online. Experiments on MiniWoB and TimeWarp show a 22.6% reduction in teacher queries and 52.1% reduction in training compute.

LLMs1 min read

AuditForecast: Structured Probabilistic Forecasting for Agents

AuditForecast is a new agentic scaffold that improves forecasting accuracy and auditability by structuring the forecasting process with explicit evidence aggregation and probability updates. It outperforms existing agentic baselines and expensive deep-research agents while maintaining a favorable cost-accuracy tradeoff.

Research1 min read

ARC-Bench: Evaluating Action Ranking in Frozen JEPA World Models

ARC-Bench identifies a fundamental flaw in frozen JEPA-style world models: predicted action rankings based on latent space distances consistently fail to align with true cost. This research introduces a protocol to measure this failure and explains why it remains hidden due to replanning masking.

PGP-Clinical-TimeKAN: Probabilistic Trajectory Forecasting - OpenSmartRoute