Skip to main content
Back to challenges
Challenge

Casualty Insurance-Linked Securities (ILS) Loss Event Trigger Monitor

Casualty ILS instruments require continuous collateral tracking and catastrophe attachment monitoring. Build a PyTorch time-series anomaly detection pipeline to evaluate cumulative casualty claims growth against ILS principal impairment triggers.

Machine LearningHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Develop a PyTorch time-series loss tracking model to monitor casualty ILS trigger conditions and collateral reserves.

Delivery guide

How work is evaluated

Evaluation

Evaluates forecasting precision of cumulative claims and trigger breach probability calculations.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

prob_bounds

Validates probability value ranges between 0.0 and 1.0

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

claims_forecasting_mae

Mean Absolute Error of predicted cumulative losses • target: 2000000 • range: 0-10000000

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Build an LSTM/Transformer time-series forecasting model in PyTorch for long-tail claims accumulation

  • Process quarterly casualty loss development statements automatically

  • Predict probability of attachment point breach over multi-year bond horizons

  • Compute capital return distributions for ILS fund investors

Resources and assets

Reference links and supporting material

Dataset notes

Quarterly casualty claims loss development time-series dataset with ILS bond tranche metadata.

Challenge setup

Evaluates forecasting precision of cumulative claims and trigger breach probability calculations.

Challenge input

JSON containing quarterly_losses array, attachment_point, and principal_amount.

PyTorch

Time-series LSTM/Transformer modeling

Evaluated output

JSON with predicted_cumulative_loss, breach_probability, and investor_principal_impairment_pct.

Evaluation checks
  • Validates probability value ranges between 0.0 and 1.0
  • Mean Absolute Error of predicted cumulative losses • target: 2000000 • range: 0-10000000
Configured evaluation evidence
  • Benchmark: CasualtyILSBench
  • Claims Forecasting Mae target: 2000000
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • Modular Mojo · Optional
  • Aim · Optional
Policy Serving
  • Qwen 3 · Required

Evaluation contract

  • prob_bounds · Weight 1
  • claims_forecasting_mae · Weight 1

Metrics

  • claims_forecasting_mae · Target 2000000

Benchmarks

  • CasualtyILSBench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing quarterly_losses array, attachment_point, and principal_amount.

    Output: JSON with predicted_cumulative_loss, breach_probability, and investor_principal_impairment_pct.

    Execution harness: Python

Evaluation tools

  • PyTorch · Deep Learning Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Casualty Insurance-Linked Securities (ILS) Loss Event Trigger Monitor