Process-Graph-FM — graph foundation model for event logs

A graph foundation model for predictive process monitoring: instead of flattening a case prefix into a sequence, it treats the prefix as a graph of events with typed directly-follows edges, and learns next-activity / next-resource prediction with a GPS-style graph transformer.

Models

EventGraphFM (model.py): GINEConv local message passing + GPSConv global attention, edge features (2), 7 numeric node features.

checkpoint	params	d_model / layers / heads	pretraining	val next-activity acc	BPIC 2015-1 zero-shot	FT test acc	FT top3	FT macro-F1
gfm-sc27/	27.28M	512 / 8 / 8	v4, 6 epochs (~0.13B positions)	0.6048	0.3102	0.4527	0.6135	0.1936
gfm-sc80/	88.14M	768 / 12 / 8	v4, 18 epochs (~0.38B positions)	0.6017	0.3226	0.4354	0.6018	0.1945

Zero-shot here is the plain global-head evaluation (no unseen-to-UNK mapping). With unseen-to-UNK: gfm-sc80 0.2975; sequence sc80 0.3810. Each folder has model.pt (state dict), config.json, vocab.json (global ids) and metrics.json.

Results (BPIC 2015-1, held out)

gfm-sc80 vs sequence sc80 — equal corpus and budget

metric	graph gfm-sc80	sequence sc80	delta
pretraining val next-activity acc	0.6017	0.6563	−5.5pp
zero-shot test acc (plain)	0.3226	0.3600	−3.7pp
zero-shot test acc (unseen-to-UNK)	0.2975	0.3810	−8.4pp
fine-tuned test acc	0.4354	0.4336	+0.0018
fine-tuned top3	0.6018	0.6280	-0.0262
fine-tuned macro-F1	0.1945	0.1861	+0.0084

Honest reading (gfm-sc80): at equal corpus and budget the 80M graph model does not beat the sequence model on zero-shot transfer, and its fine-tuned accuracy is higher than sc80 (+0.0018). Its pretraining validation is also lower. The fine-tuning required a lower learning rate (1e-4 instead of 3e-4) because the 80M graph diverged at the recipe used for sc27/sc80; this is documented rather than hidden. One seed, under-budget: the result is evidence, not a graph-vs-sequence law.

gfm-sc27 vs sequence sc27 — equal corpus and budget

metric	graph gfm-sc27	sequence sc27	delta
pretraining val next-activity acc	0.6048	0.6451	−4.0pp
zero-shot test acc	0.3102	0.3068	+0.34pp
fine-tuned test acc	0.4527	0.4296	+2.3pp
fine-tuned top3	0.6135	0.5865	+2.7pp
fine-tuned macro-F1	0.1936	0.1877	+0.6pp

Late fusion graph + sequence (validation-selected)

seq=0.4336 graph=0.4354 best_single=0.4354 mix_sc80_gfm80.json: weights=seq=0.60 graph=0.40 val={'accuracy': 0.4945067391550572, 'top3': 0.6815041341035225, 'macro_f1': 0.25174367057122143} test={'accuracy': 0.46706586826347307, 'top3': 0.6551183347590533, 'macro_f1': 0.21542101104609862} beat_best_single_by=+0.0317 gate=PASS mix_sc80_gfm80_v6.json: weights=seq=0.40 graph=0.35 v6=0.25 val={'accuracy': 0.507305470608223, 'top3': 0.691924340242383, 'macro_f1': 0.271597143329999} test={'accuracy': 0.46706586826347307, 'top3': 0.6515540347875677, 'macro_f1': 0.2305893787932908} beat_best_single_by=+0.0317 gate=PASS

The fusion script combines the fine-tuned probability matrices of sc80 and gfm-sc80 with convex weights selected on the validation split, then reports the untouched test split. Full grid and both validation criteria are in out/mix_sc80_gfm80.json of the study repository; a 3-way variant including fm-v6 is in out/mix_sc80_gfm80_v6.json.

Usage

import json
import torch
from model import EventGraphFM

cfg = json.load(open("gfm-sc80/config.json"))
state = torch.load("gfm-sc80/model.pt", map_location="cpu", weights_only=True)
model = EventGraphFM(
    n_act=cfg["n_act"], n_res=cfg["n_res"], n_logs=cfg["n_logs"],
    d_model=cfg["d_model"], n_layers=cfg["n_layers"], n_heads=cfg["n_heads"],
    dropout=cfg["dropout"], edge_dim=cfg["edge_dim"], num_dim=cfg["num_dim"],
)
model.load_state_dict(state)
model.eval()

# forward(gact, gres, num, log_id, edge_index, edge_attr, batch_vec, mask_flag)
# gact/gres: global ids (activities 4..776, resources 2..2739, UNK=1)
# edge_index/edge_attr: typed directly-follows graph; build it with the project code

The graph construction (DFG edges, edge types, numeric node features, masking) follows the ProcessGraph-FM project code. The global activity/resource ids are in gfm-sc80/vocab.json; log ids are conditioning inputs.

Data & licensing

Weights and code are released under MIT. Raw event logs are not redistributed and keep their original licenses (mostly 4TU General Terms of Use; BPIC 2018 is CC0). The synthetic half of the corpus is available at Polifemus/event-log-mixer under MIT.

Related repositories

• Polifemus/process-fm — sequence foundation models and scaling suite.
• Polifemus/prefix-next-activity — GNN vs fine-tuned transformer study.
• Polifemus/event-log-mixer — synthetic event-log corpus.

Citation

@misc{process-graph-fm,
  title  = {Process-Graph-FM: A GPS Graph Foundation Model for Event Logs},
  author = {Polidori, Filippo},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/Polifemus/process-graph-fm}}
}