# Neural Unification for Logic Reasoning over Natural Language

arXiv: https://arxiv.org/abs/2109.08460 · HTML: https://mlsanity.com/paper/2109.08460/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 32 references that are in the map.
- First posted on arXiv 17 Sep 2021.

## Builds on (most cited first, 20 of 20)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Teaching Machines to Read and Comprehend](https://mlsanity.com/paper/1506.03340/) (2015)
- [SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference](https://mlsanity.com/paper/1808.05326/) (2018)
- [ProofWriter: Generating Implications, Proofs, and Abductive Statements over Natural Language](https://mlsanity.com/paper/2012.13048/) (2020)
- [Transformers as Soft Reasoners over Language](https://mlsanity.com/paper/2002.05867/) (2020)
- [Compositional Attention Networks for Machine Reasoning](https://mlsanity.com/paper/1803.03067/) (2018)
- [End-to-End Differentiable Proving](https://mlsanity.com/paper/1705.11040/) (2017)
- [Neural Logic Machines](https://mlsanity.com/paper/1904.11694/) (2019)
- [PRover: Proof Generation for Interpretable Reasoning over Rules](https://mlsanity.com/paper/2010.02830/) (2020)
- [Differentiable Reasoning over a Virtual Knowledge Base](https://mlsanity.com/paper/2002.10640/) (2020)
- [NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language](https://mlsanity.com/paper/1906.06187/) (2019)
- [Differentiable Reasoning on Large Knowledge Bases and Natural Language](https://mlsanity.com/paper/1912.10824/) (2019)
- [Exploring Neural Models for Parsing Natural Language into First-Order Logic](https://mlsanity.com/paper/2002.06544/) (2020)
- [The fundamental nature of the log loss function](https://mlsanity.com/paper/1502.06254/) (2015)
- [A Survey on Theorem Provers in Formal Methods](https://mlsanity.com/paper/1912.03028/) (2019)
- [Deep Learning for Ontology Reasoning](https://mlsanity.com/paper/1705.10342/) (2017)
- [Towards a Model Theory for Distributed Representations](https://mlsanity.com/paper/1410.5859/) (2014)

## Built on by (most cited first, 8 of 8)

- [State-of-the-art generalisation research in NLP: A taxonomy and review](https://mlsanity.com/paper/2210.03050/) (2022)
- [A Comprehensive Survey on Applications of Transformers for Deep Learning Tasks](https://mlsanity.com/paper/2306.07303/) (2023)
- [Learning Deductive Reasoning from Synthetic Corpus based on Formal Logic](https://mlsanity.com/paper/2308.07336/) (2023)
- [NELLIE: A Neuro-Symbolic Inference Engine for Grounded, Compositional, and Explainable Reasoning](https://mlsanity.com/paper/2209.07662/) (2022)
- [Explicit Planning Helps Language Models in Logical Reasoning](https://mlsanity.com/paper/2303.15714/) (2023)
- [Empirical Investigation of Neural Symbolic Reasoning Strategies](https://mlsanity.com/paper/2302.08148/) (2023)
- [Semantic-aware Dynamic Retrospective-Prospective Reasoning for Event-level Video Question Answering](https://mlsanity.com/paper/2305.08059/) (2023)
- [LLM Guided Inductive Inference for Solving Compositional Problems](https://mlsanity.com/paper/2309.11688/) (2023)

## Abstract (arXiv)

Automated Theorem Proving (ATP) deals with the development of computer programs being able to show that some conjectures (queries) are a logical consequence of a set of axioms (facts and rules). There exists several successful ATPs where conjectures and axioms are formally provided (e.g. formalised as First Order Logic formulas). Recent approaches, such as (Clark et al., 2020), have proposed transformer-based architectures for deriving conjectures given axioms expressed in natural language (English). The conjecture is verified through a binary text classifier, where the transformers model is trained to predict the truth value of a conjecture given the axioms. The RuleTaker approach of (Clark et al., 2020) achieves appealing results both in terms of accuracy and in the ability to generalize, showing that when the model is trained with deep enough queries (at least 3 inference steps), the transformers are able to correctly answer the majority of queries (97.6%) that require up to 5 inference steps. In this work we propose a new architecture, namely the Neural Unifier, and a relative training procedure, which achieves state-of-the-art results in term of generalisation, showing that mimicking a well-known inference procedure, the backward chaining, it is possible to answer deep queries even when the model is trained only on shallow ones. The approach is demonstrated in experiments using a diverse set of benchmark data.