Syntactic Language Models (SLMs) can be trained efficiently to reach relatively high performance; however, they have trouble with inference efficiency due to the explicit generation of syntactic structures. In this paper, we propose a new method dubbed tree-planting: instead of explicitly generating syntactic structures, we "plant" trees into attention weights of unidirectional Transformer LMs to implicitly reflect syntactic structures of natural language. Specifically, unidirectional Transformer LMs trained with tree-planting will be called Tree-Planted Transformers (TPT), which inherit the training efficiency from SLMs without changing the inference efficiency of their underlying Transformer LMs. Targeted syntactic evaluations on the SyntaxGym benchmark demonstrated that TPTs, despite the lack of explicit generation of syntactic structures, significantly outperformed not only vanilla Transformer LMs but also various SLMs that generate hundreds of syntactic structures in parallel. This result suggests that TPTs can learn human-like syntactic knowledge as data-efficiently as SLMs while maintaining the modeling space of Transformer LMs unchanged.

Radar sensors are low cost, long-range, and weather-resilient. Therefore, they are widely used for driver assistance functions, and are expected to be crucial for the success of autonomous driving in the future. In many perception tasks only pre-processed radar point clouds are considered. In contrast, radar spectra are a raw form of radar measurements and contain more information than radar point clouds. However, radar spectra are rather difficult to interpret. In this work, we aim to explore the semantic information contained in spectra in the context of automated driving, thereby moving towards better interpretability of radar spectra. To this end, we create a radar spectra-language model, allowing us to query radar spectra measurements for the presence of scene elements using free text. We overcome the scarcity of radar spectra data by matching the embedding space of an existing vision-language model (VLM). Finally, we explore the benefit of the learned representation for scene parsing, and obtain improvements in free space segmentation and object detection merely by injecting the spectra embedding into a baseline model.

TheBench is a tool to study monadic structures in natural language. It is for writing monadic grammars to explore analyses, compare diverse languages through their categories, and to train models of grammar from form-meaning pairs where syntax is latent variable. Monadic structures are binary combinations of elements that employ semantics of composition only. TheBench is essentially old-school categorial grammar to syntacticize the idea, with the implication that although syntax is autonomous (recall \emph{colorless green ideas sleep furiously}), the treasure is in the baggage it carries at every step, viz. semantics, more narrowly, predicate-argument structures indicating choice of categorial reference and its consequent placeholders for decision in such structures. There is some new thought in old school. Unlike traditional categorial grammars, application is turned into composition in monadic analysis. Moreover, every correspondence requires specifying two command relations, one on syntactic command and the other on semantic command. A monadic grammar of TheBench contains only synthetic elements (called `objects' in category theory of mathematics) that are shaped by this analytic invariant, viz. composition. Both ingredients (command relations) of any analytic step must therefore be functions (`arrows' in category theory). TheBench is one implementation of the idea for iterative development of such functions along with grammar of synthetic elements.

Large scale neural models show impressive performance across a wide array of linguistic tasks. Despite this they remain, largely, black-boxes - inducing vector-representations of their input that prove difficult to interpret. This limits our ability to understand what they learn, and when the learn it, or describe what kinds of representations generalise well out of distribution. To address this we introduce a novel approach to interpretability that looks at the mapping a model learns from sentences to representations as a kind of language in its own right. In doing so we introduce a set of information-theoretic measures that quantify how structured a model's representations are with respect to its input, and when during training that structure arises. Our measures are fast to compute, grounded in linguistic theory, and can predict which models will generalise best based on their representations. We use these measures to describe two distinct phases of training a transformer: an initial phase of in-distribution learning which reduces task loss, then a second stage where representations becoming robust to noise. Generalisation performance begins to increase during this second phase, drawing a link between generalisation and robustness to noise. Finally we look at how model size affects the structure of the representational space, showing that larger models ultimately compress their representations more than their smaller counterparts.

Discourse Representation Structure (DRS) is an innovative semantic representation designed to capture the meaning of texts with arbitrary lengths across languages. The semantic representation parsing is essential for achieving natural language understanding through logical forms. Nevertheless, the performance of DRS parsing models remains constrained when trained exclusively on monolingual data. To tackle this issue, we introduce a cross-lingual training strategy. The proposed method is model-agnostic yet highly effective. It leverages cross-lingual training data and fully exploits the alignments between languages encoded in pre-trained language models. The experiments conducted on the standard benchmarks demonstrate that models trained using the cross-lingual training method exhibit significant improvements in DRS clause and graph parsing in English, German, Italian and Dutch. Comparing our final models to previous works, we achieve state-of-the-art results in the standard benchmarks. Furthermore, the detailed analysis provides deep insights into the performance of the parsers, offering inspiration for future research in DRS parsing. We keep updating new results on benchmarks to the appendix.

This paper revisits recent code similarity evaluation metrics, particularly focusing on the application of Abstract Syntax Tree (AST) editing distance in diverse programming languages. In particular, we explore the usefulness of these metrics and compare them to traditional sequence similarity metrics. Our experiments showcase the effectiveness of AST editing distance in capturing intricate code structures, revealing a high correlation with established metrics. Furthermore, we explore the strengths and weaknesses of AST editing distance and prompt-based GPT similarity scores in comparison to BLEU score, execution match, and Jaccard Similarity. We propose, optimize, and publish an adaptable metric that demonstrates effectiveness across all tested languages, representing an enhanced version of Tree Similarity of Edit Distance (TSED).

Incremental models that process sentences one token at a time will sometimes encounter points where more than one interpretation is possible. Causal models are forced to output one interpretation and continue, whereas models that can revise may edit their previous output as the ambiguity is resolved. In this work, we look at how restart-incremental Transformers build and update internal states, in an effort to shed light on what processes cause revisions not viable in autoregressive models. We propose an interpretable way to analyse the incremental states, showing that their sequential structure encodes information on the garden path effect and its resolution. Our method brings insights on various bidirectional encoders for contextualised meaning representation and dependency parsing, contributing to show their advantage over causal models when it comes to revisions.

This study explores the formality style transfer in Persian, particularly relevant in the face of the increasing prevalence of informal language on digital platforms, which poses challenges for existing Natural Language Processing (NLP) tools. The aim is to transform informal text into formal while retaining the original meaning, addressing both lexical and syntactic differences. We introduce a novel model, Fa-BERT2BERT, based on the Fa-BERT architecture, incorporating consistency learning and gradient-based dynamic weighting. This approach improves the model's understanding of syntactic variations, balancing loss components effectively during training. Our evaluation of Fa-BERT2BERT against existing methods employs new metrics designed to accurately measure syntactic and stylistic changes. Results demonstrate our model's superior performance over traditional techniques across various metrics, including BLEU, BERT score, Rouge-l, and proposed metrics underscoring its ability to adeptly navigate the complexities of Persian language style transfer. This study significantly contributes to Persian language processing by enhancing the accuracy and functionality of NLP models and thereby supports the development of more efficient and reliable NLP applications, capable of handling language style transformation effectively, thereby streamlining content moderation, enhancing data mining results, and facilitating cross-cultural communication.

We introduce SPAGHETTI: Semantic Parsing Augmented Generation for Hybrid English information from Text Tables and Infoboxes, a hybrid question-answering (QA) pipeline that utilizes information from heterogeneous knowledge sources, including knowledge base, text, tables, and infoboxes. Our LLM-augmented approach achieves state-of-the-art performance on the Compmix dataset, the most comprehensive heterogeneous open-domain QA dataset, with 56.5% exact match (EM) rate. More importantly, manual analysis on a sample of the dataset suggests that SPAGHETTI is more than 90% accurate, indicating that EM is no longer suitable for assessing the capabilities of QA systems today.

Natural language is structured hierarchically: words are grouped into phrases or constituents, which can be further grouped to form higher-level phrases up to the full sentence. How well do the neural network models trained on language data learn this phrase structure of human language has been a subject of great interest. A flurry of past work have shown that syntax trees can be recovered from recurrent neural network (RNN) and transformer-based models trained on large-scale language corpora (Tenney et al., 2019, Peters et al., 2018, Lin et al., 2019, Wu et al., 2020). While these studies provide useful evidence of the aforementioned phenomenon, they do not shed light on the architectural choices, training paradigms or dataset characteristics that lead models to learn the phrase structure of language. A useful tool to understand these model and dataset specific properties is through the test for hierarchical generalization, i.e., evaluating the capability of a model to generalize to novel syntactic forms, which were unseen during training. A classic problem to test for hierarchical generalization is question formation, where given a declarative sentence, e.g., My walrus does move the dogs that do wait., the task is to transform it into a question: Does my walrus move the dogs that do wait? The task is accomplished by moving one auxiliary verb to the front. The correct choice to move does in this example (rather than do), is predicted both by a hierarchical rule based on the phrase-structure syntax of the sentence, and by a linear rule that says to move the first auxiliary. Hence, as a test for hierarchical generalization, we can ask, for neural networks trained from scratch on data that is consistent with both hierarchical and linear rules (i.e.,