Payment fraud causes financial losses and undermines users' trust in web payment services. Although fraud detection is a critical defense for payment security, complex and diverse fraudulent behaviors can still evade detection. Generating fraud samples for data augmentation can broaden detectors' coverage of different fraud patterns. However, existing methods matching the distribution of transaction attributes alone do not ensure that generated sequences preserve the sequential relationships among transaction choices, payment feedback, and subsequent behavioral adjustments. To this end, we propose PaySeqGen, an interpretable inverse reinforcement learning framework that reconstructs fraud behavior policies from fraudsters' observed payment trajectories. We design a mechanism for constructing candidate actions to support plausible payment sequence generation while enabling exploration. The framework also learns interpretable behavioral rewards that capture fraudsters' preferences across different transaction contexts. By alternating between reward learning and policy optimization, the framework reconstructs policies that generate new fraud sequences for detector augmentation and learns interpretable rewards supporting sequential fraud-pattern analysis in real trajectories. Experiments in real-world payment scenarios at Ant International, a leading global digital payment provider, show that PaySeqGen outperforms state-of-the-art generative augmentation methods. Reward-based trajectory analysis also identifies fraud patterns of practical value, helping analyst locating suspicious segments and examine multistep fraud tactics. Evaluation in the deployed system of online payment further confirms its effectiveness.
Figures & tables
Figure 1. An illustrative online fraud payment sequence and its decision context shaped by feedback.
Figure 2. Overview of PaySeqGen. (a) Model payment trajectories as a sequential decision process. (b) Construct candidate actions through history-guided sampling and exploratory sampling. (c) Learn behavioral rewards with feedback–action interactions, supporting fraud-pattern analysis and guiding fraud policy reconstruction. (d) Optimize the policy using learned rewards to generate payment sequences for detector augmentation.
Detectors
RT
TabFormer
GraphSAGE
XGBoost
Metrics
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Control
Fraud replay
-0.046
+1.094
+0.166
-0.092
-1.679
-0.042
+0.036
+1.098
+0.416
+0.109
+0.325
+0.074
Baselines
BC
-0.141
-0.360
+0.055
+0.089
-2.972
-0.097
+0.681
+3.375
+0.975
-0.195
-1.266
-0.354
CTGAN
-0.036
+1.559
-0.169
+0.148
-2.218
+0.170
+0.066
+1.370
+0.934
-0.025
-0.193
-0.304
Diffusion-TS
-0.189
+0.871
-0.610
-0.042
-1.923
-0.423
+0.343
+1.454
+1.056
-0.095
-0.976
-0.232
TabDiff
+0.169
+0.769
+0.108
+0.198
-2.000
+0.295
-0.251
+0.111
-0.517
-0.093
-0.103
-0.628
Table 1. Performance changes across four fraud detectors under different augmentation methods. Values are percentage-point differences from training without augmentation. Boldface and underlining indicate the best and second-best results per column.
Figure 3. Reward-guided fraud-pattern analysis. (a) Payment history and relative rewards, with the selected interval shaded and the fraudulent target marked by a star. (b) Features with the largest positive contributions to relative reward within the selected interval. (c) Prevalence of the corresponding pattern among real fraud-target and legitimate-target trajectories.
Figure 4. Sensitivity Analysis with respect to the ratio of augmented samples and the exploration candidate ratio.
Figure 5. Schematic overview of online payment risk-control pipeline integrating risk services and model services.
Method
AUC
Recall@2
KS
Model-1
86.02
40.66
54.69
Model-1 (PaySeqGen)
86.23
43.85
55.45
Model-2
85.90
41.79
55.03
Model-2 (PaySeqGen)
86.16
44.63
56.09
Table 2. Results of online A/B tests.
Appendix figures & tables3 assets
Supplementary material from the paper’s appendix.
Appendix
Augmentation
RT
TabFormer
GraphSAGE
XGBoost
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Δ AUC
Δ Recall@2
Δ KS
Real-fraud replay
-3.874
-0.654
-2.781
+0.249
-0.292
+0.159
-1.221
-0.905
-0.327
-0.874
-0.028
-0.242
BC
-11.970
-0.710
-10.376
-1.583
-0.306
-1.531
-0.439
-1.184
+0.708
-1.175
+0.084
-1.019
CTGAN
-12.406
-1.058
-11.004
-4.368
-0.794
-3.842
-6.084
-1.351
-5.278
-0.404
+0.292
-0.234
Diffusion-TS
-4.191
-0.947
-2.809
+0.449
-0.111
+0.559
-0.548
-1.253
+0.258
-0.894
+0.014
-0.401
TabDiff
-2.361
-0.961
-3.202
-4.600
-0.473
-4.566
-3.620
-1.392
-3.241
-0.337
+0.265
-0.150
Appendix
Table 3. Augmentation results on public credit card fraud detection dataset MS-FFSD.
Reduction
Fraud-target
Legitimate-target
10%
43/885 (4.86%)
5/1,604 (0.31%)
20%
35/885 (3.95%)
4/1,604 (0.25%)
30%
31/885 (3.50%)
3/1,604 (0.19%)
Appendix
Table 4. Pattern prevalence under different minimum amount reductions. Entries show matched/eligible trajectories and occurrence rates.
Method
Local W1 ↓
Cond. JS ↓
2-step JS ↓
Path MMD 2 ↓
Cov. (%) ↑
BC
0.316
0.180
0.486
0.0313
24.44
CTGAN
1.462
0.380
0.349
0.2352
1.39
Diffusion-TS
0.613
0.299
0.284
0.0574
10.21
TabDiff
1.086
0.335
0.240
0.0373
0.98
MoE-BiEntIRL
0.730
0.451
0.472
0.1407
4.97
PaySeqGen
0.560
0.369
0.176
0.0540
18.12
Appendix
Table 5. Behavioral fidelity and coverage of generated fraud samples on the proprietary dataset. Values are means; bold denotes the best value in each column.
Real-time payment fraud detection is a non-stationary streaming prediction problem: adversaries adapt before supervised labels mature, and localized burst attacks can cause losses before retraining. Production systems typically rely on tabular classifiers and rules, which can struggle to capture these emerging sequential patterns before periodic retraining occurs. We present SR-Fraud, an outcome-supervised reflective LLM framework that decouples request-time decisions from offline adaptation. A frozen, stateless agent scores each transaction from a Hybrid Episodic Window to track behavioral shifts, while an offline reflection agent proposes boundary hypotheses from matured errors. A deterministic verifier then admits only supported hypotheses into an executable knowledge state. On a production payment-fraud benchmark, SR-Fraud improves all detection metrics over its frozen decision agent, obtains higher point estimates than static and periodically retrained CatBoost, and detects an emerging fraud burst.
Production fraud-detection models learn from a censored slice of their own traffic: transactions a legacy system blocked never receive fraud labels, so models train on a biased, approved-only distribution. We extend LLM-guided reward design, the framework of Qu et al. (2025) in which an LLM evolves the reward of a two-step REINFORCE fraud-blocking policy, to attack this reject-inference problem inside the reward. A single LLM designs a dual-stream reward: stream 1 applies the standard dollar-weighted precision objective to labeled transactions; stream 2 scores legacy-blocked, unlabeled transactions by contrasting the newest gradient-boosted risk signal against the legacy scores that drove the original block. Rather than prescribe which signals matter, we expose a superset of evidence and let the LLM decide for itself which to use, guided by evaluation feedback alone; because it writes the reward as code, that choice is directly auditable. On a held-out four-month out-of-time test, the dual-stream agent beats the prior labeled-only agent at every target recall, by +4.7, +2.4 and +3.0 points of dollar-weighted precision at 80%, 85% and 90% recall, and the production baseline by +6.8 points on average. The resulting policy is architecturally identical to that of Qu et al. (2025) and adds no inference cost.
Ant International accumulates large volumes of transaction history across users and payment businesses. Confirmed fraud and chargeback outcomes, by contrast, are scarce, delayed, and specific to each business. A single encoder pretrained on unlabeled histories could serve all businesses, but only if pretraining captures the behavior that separates risk. Coarse activity differences can make contrastive pretraining on transaction logs too easy: transaction volume, amount, and timing can distinguish a history's continuation from arbitrary candidates without modeling how payment resources are used. We present TRIDENT-Seq, a transaction foundation model trained with Prefix-Future Contrast (PFC). PFC matches each prefix to a summary of its subsequent behavior among candidates with similar activity profiles; activity prediction and dispersion regularization act on the encoder output. Across three production payment businesses, the pretrained encoder improves average precision (AP) and recall at a 0.5% intervention budget over architecture-matched scratch training in all six business-budget settings. With a fixed validation set, it exceeds full-label scratch training in mean AP and recall using 10% of the training labels on Wallet and 25% on Shopping and SME. Pretraining only on another business's histories still helps. Replacing profile-matched candidates with random ones cuts recall from 24.62% to 5.64%. In Ant International's deployment evaluation, recorded amount-based chargeback and overall loss rates are 37.7% and 37.9% lower than those of the previous XGBoost-based system under the same intervention budget.
Weiyu Han, Haoyu Jiang, Jie Sheng +7
Ant International · East China Normal University · Tongji University +1