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.
Figures & tables
Figure 1. TRIDENT-Seq overview. (a) PFC learns behavioral representations without risk labels. (b) The shared foundation model adapts to each business through supervised fine-tuning (SFT) with limited labels. (c) At Ant International, online A/B evaluation across deployed payment businesses shows a 37.9% reduction in the overall loss rate under the same intervention budget. Three stages: behavior pretraining from prefixes and future summaries; reuse of the same pretrained encoder for business-specific supervised fine-tuning; and production risk evaluation.
Figure 2. TRIDENT-Seq architecture. (a) Event tokenization. (b) Shared history encoder. (c) Pretraining with PFC, activity prediction, and dispersion regularization, and business-specific supervised adaptation. Event attributes are encoded into tokens. A Transformer produces a history representation. Pretraining uses prefix and summary projection heads; downstream adaptation uses a business risk head.
Table 2. Performance comparison (%). Best results are bold; second-best results are underlined.
Figure 3. Label efficiency on Shopping and SME. Four plots show AP and recall at a 0.5 percent intervention budget for TRIDENT-Seq and scratch training across five training-label budgets on Shopping and SME, with a dashed line for full-label scratch.
Table 4. Wallet ablation: recall (%) at three intervention budgets.
Figure 4. Pretrained representations of annotated histories. PCA and t-SNE projections of pretrained representations, with two behavior groups and an outlined t-SNE region used for case inspection.
Figure 5. Card and device switching after rejection. Three payment attempts show changes in anonymous card and device identities following rejections while retaining the IP identity, ending in a successful attempt.
Method
Traffic (M)
CB rate change ↓
Loss rate change ↓
XGB-AutoRefit
9.06
base
base
SAINT-style
10.56
− 13.21%
− 18.18%
TRIDENT-Seq
8.58
− 37.74%
− 37.88%
Table 5. Portfolio-level online A/B results at Ant International. Pooled arm-level rate changes relative to XGB.
Appendix figures & tables12 assets
Supplementary material from the paper’s appendix.
Appendix
Setting
Value
Layers / hidden size / heads
8 / 1,024 / 8
Encoder parameters / history cap
92.92M / 300 events
Hash bits / probes
128 / 4 per field
Eligible history / prefix
At least 2 events / older 60%
Future summary
24 dimensions (Table 7 )
Profile
Three coordinates (Equation 4 )
Appendix
Table 6. Encoder and pretraining configuration.
Coordinates
Definition
1
Constant 1
2
min(10,Std(Δδ)/max(Mean∣Δδ∣,10−6))/10
3–5
Mean(u)/16 , Std(u)/16 , max(u)/16
6–7
Mean(δ)/92 , Std(δ)/92
8
∑j,khjk/(1536m)
9
Field-level switching rate SW between adjacent events
Acquiring fraud labels for payment transactions is expensive and inconsistent across tasks, and high-cardinality identifiers such as merchant names make one-hot feature pipelines infeasible at scale. We present CARTA, a self-supervised encoder that learns a compact 16-dim embedding from raw transactions without labels; the frozen embedding is then reused for label-efficient fraud probing and exploratory unsupervised risk stratification without retraining the encoder. CARTA uses type-specific projections with learned embedding tables for high-cardinality identifiers, trained with token excision and InfoNCE contrastive loss. On IEEE-CIS, its linear probe at 1% of labels exceeds a raw-feature probe at 100%; on IBM TabFormer (2M rows), its frozen embedding reaches GBDT AUPRC 0.508, above frequency encoding (0.411) and the merchant-dropping baseline (0.022), though fully-supervised CatBoost is stronger. CARTA is a label-free, reusable representation that complements task-specific supervised models.
Shreyas K. Chandrahas, Jose Munoz, Satish Babu Bellapu
Visa Inc. Foster City, CA, USA · Massachusetts Institute of Technology Cambridge, MA, USA
Merchant risk control at large payment platforms screens tens of millions of merchants daily, where false positives harm legitimate merchants and false negatives leave harmful activity undetected. The hardest cases require jointly understanding a merchant's textual profile and long behavioral sequence. Large language models (LLMs) excel at text but cannot natively model such sequences, while adapting them often causes catastrophic forgetting. We present SeqLLM, a framework that adds behavioral-sequence modeling to a pretrained LLM while preserving its language ability. SeqLLM combines three components: a compact discrete vocabulary that represents behavioral events as native tokens; a lightweight projector, trained with a two-stage alignment curriculum, that grounds these tokens in the LLM's semantic space; and prefix-guided capability injection, which acquires sequence-modeling ability through task-prefixed supervised fine-tuning rather than continual pre-training. SeqLLM is deployed at WeChat Pay, screening millions of merchants daily. Against the production DeepSeek-based LLM baseline, it raises screening precision from 92.0% to 97.5%. Its pretrained behavior-token embeddings also improve [email protected]% by 26.8 percentage points in a production fraud detector serving billion-scale transaction traffic. Beyond payments, SeqLLM achieves state-of-the-art results on public recommendation benchmarks. On MovieLens and Amazon, it surpasses the strong User-LLM baseline by up to 32% relative Recall@5 while retaining markedly stronger language ability. On RecIF, it improves Pass@32 by 14.2% over the full OneRec-8B pipeline using only one-fifth of its GPU-days.
Guilin Li, Jiaxing Zhang, Matthias Hwai Yong Tan +2
Fraud detection models in payment networks train on chargeback labels that are systematically biased. Every label must survive three sequential gates: authorization (declined transactions generate no labels), issuer reporting (unreported fraud is invisible), and delay (pending chargebacks are missing at training time). Labels that do arrive may be corrupted by first-party misuse or issuer misclassification. A companion paper [arXiv:2605.27557] proved that these four impairments impose a minimax lower bound on detection performance. This paper asks: can that bound be achieved? We formalize the observation pipeline as a sequential missing-data problem with three propensity stages and a corruption layer, and construct the Sequential Triply Robust (STR) estimator. The STR corrects for all four impairments simultaneously and achieves the semiparametric efficiency bound -- no estimator can have lower asymptotic variance. It is sequentially triply robust: at each gate, consistency requires only that either the propensity model or the outcome regression is correctly specified, not both. We provide corruption correction via noise-rate-adjusted pseudo-labels, empirical Bayes shrinkage to stabilize inverse-propensity weights for small issuers, a plug-in variance estimator yielding valid confidence intervals, and a Bernstein concentration inequality for finite-sample guarantees. On the operational side, we derive the optimal training delay -- the maturity window that minimizes the sum of label-quality loss and model staleness -- and prove that the STR permits training on data that is days old rather than months old, decoupling model freshness from the chargeback maturity cycle. The STR provably dominates naive chargeback-based training in mean squared error for any sample size.