Papers › ORCHARD: A Benchmark For Measuring Systematic Generalization of Multi-Hierarchical Reasoning

ORCHARD: A Benchmark For Measuring Systematic Generalization of Multi-Hierarchical Reasoning

28 Nov 2021arXiv:2111.14034archive 2025-07-28

Bill Tuck Weng Pung, Alvin Chan

The ability to reason with multiple hierarchical structures is an attractive and desirable property of sequential inductive biases for natural language processing. Do the state-of-the-art Transformers and LSTM architectures implicitly encode for these biases? To answer this, we propose ORCHARD, a diagnostic dataset for systematically evaluating hierarchical reasoning in state-of-the-art neural sequence models. While there have been prior evaluation frameworks such as ListOps or Logical Inference, our work presents a novel and more natural setting where our models learn to reason with multiple explicit hierarchical structures instead of only one, i.e., requiring the ability to do both long-term sequence memorizing, relational reasoning while reasoning with hierarchical structure. Consequently, backed by a set of rigorous experiments, we show that (1) Transformer and LSTM models surprisingly fail in systematic generalization, and (2) with increased references between hierarchies, Transformer performs no better than random.

PaperPDFCode

Code

billptw/orchard officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

DiagnosticListOpsRelational ReasoningSystematic Generalization

Datasets

Introduced by this paper, per the archive.

Orchard

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLSTMLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSigmoid ActivationSoftmaxTanh ActivationTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections