# Learning Proposition-Level Links from Sentence-Level Supervision for Neuro-Symbolic NLI

arXiv: https://arxiv.org/abs/2610.15158 · HTML: https://mlsanity.com/paper/2610.15158/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 51 references that are in the map.
- First posted on arXiv 10 Oct 2026.

## Builds on (most cited first, 23 of 23)

- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [A large annotated corpus for learning natural language inference](https://mlsanity.com/paper/1508.05326/) (2015)
- [MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers](https://mlsanity.com/paper/2002.10957/) (2020)
- [DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing](https://mlsanity.com/paper/2111.09543/) (2021)
- [C-Pack: Packed Resources For General Chinese Embeddings](https://mlsanity.com/paper/2309.07597/) (2023)
- [A Decomposable Attention Model for Natural Language Inference](https://mlsanity.com/paper/1606.01933/) (2016)
- [Reasoning about Entailment with Neural Attention](https://mlsanity.com/paper/1509.06664/) (2015)
- [Maximum Bayes Smatch Ensemble Distillation for AMR Parsing](https://mlsanity.com/paper/2112.07790/) (2021)
- [NeuralLog: Natural Language Inference with Joint Neural and Logical Reasoning](https://mlsanity.com/paper/2105.14167/) (2021)
- [MonaLog: a Lightweight System for Natural Language Inference Based on Monotonicity](https://mlsanity.com/paper/1910.08772/) (2019)
- [Logical Reasoning with Span-Level Predictions for Interpretable and Robust NLI Models](https://mlsanity.com/paper/2205.11432/) (2022)
- [Atomic Inference for NLI with Generated Facts as Atoms](https://mlsanity.com/paper/2305.13214/) (2023)
- [NLI under the Microscope: What Atomic Hypothesis Decomposition Reveals](https://mlsanity.com/paper/2502.08080/) (2025)
- [Exploring End-to-End Differentiable Natural Logic Modeling](https://mlsanity.com/paper/2011.04044/) (2020)
- [Weakly Supervised Explainable Phrasal Reasoning with Neural Fuzzy Logic](https://mlsanity.com/paper/2109.08927/) (2021)
- [Acquisition of Phrase Correspondences using Natural Deduction Proofs](https://mlsanity.com/paper/1804.07656/) (2018)
- [Representing Meaning with a Combination of Logical and Distributional Models](https://mlsanity.com/paper/1505.06816/) (2015)
- [Entailment-Preserving First-order Logic Representations in Natural Language Entailment](https://mlsanity.com/paper/2502.16757/) (2025)
- [Can LLMs Interpret and Leverage Structured Linguistic Representations? A Case Study with AMRs](https://mlsanity.com/paper/2504.04745/) (2025)
- [Extractive Fact Decomposition for Interpretable Natural Language Inference in one Forward Pass](https://mlsanity.com/paper/2509.18901/) (2025)
- [Atomic-SNLI: Fine-Grained Natural Language Inference through Atomic Fact Decomposition](https://mlsanity.com/paper/2601.06528/) (2026)
- [When Does Meaning Backfire? Investigating the Role of AMRs in NLI](https://mlsanity.com/paper/2506.14613/) (2025)
- [Pairwise Logical Selection of Enthymeme Completions under Semantic-Link Uncertainty](https://mlsanity.com/paper/2608.18820/) (2026)

## Abstract (arXiv)

We study neuro-symbolic natural language inference (NLI) with independently translated propositional representations of the premise and hypothesis. Before symbolic reasoning, the system must infer semantic links between their proposition-level units (atoms). Existing systems based on Abstract Meaning Representation (AMR) and Boolean satisfiability (SAT) identify these links using fixed embedding-based source selection and pretrained sentence-level NLI models not trained for atom-level linking. We introduce a trainable atom linker that jointly predicts link existence, source selection, and relation type. Without atom-pair training annotations, we treat cross-formula link structures as latent and train the linker from sentence-level NLI labels, keeping the translator and SAT reasoner fixed. We evaluate on the full public test splits of Stanford Natural Language Inference (SNLI) and Sentences Involving Compositional Knowledge (SICK), and the matched and mismatched validation splits of Multi-Genre Natural Language Inference (MNLI). Our system improves accuracy by 11.40-27.09 percentage points over a baseline applying sentence-trained NLI models unchanged to atom pairs within the same SAT pipeline. It also improves over the prior fixed-linker AMR-to-SAT system by 8.14-26.32 percentage points on SNLI and both MNLI splits, with a smaller gain on SICK.