Milan Gritta

Songbo Hu, Han Zhou, Mete Hergul et al. · cambridge

Creating high-quality annotated data for task-oriented dialog (ToD) is known to be notoriously difficult, and the challenges are amplified when the goal is to create equitable, culturally adapted, and large-scale ToD datasets for multiple languages. Therefore, the current datasets are still very scarce and suffer from limitations such as translation-based non-native dialogs with translation artefacts, small scale, or lack of cultural adaptation, among others. In this work, we first take stock of the current landscape of multilingual ToD datasets, offering a systematic overview of their properties and limitations. Aiming to reduce all the detected limitations, we then introduce Multi3WOZ, a novel multilingual, multi-domain, multi-parallel ToD dataset. It is large-scale and offers culturally adapted dialogs in 4 languages to enable training and evaluation of multilingual and cross-lingual ToD systems. We describe a complex bottom-up data collection process that yielded the final dataset, and offer the first sets of baseline scores across different ToD-related tasks for future reference, also highlighting its challenging nature.

Fenia Christopoulou, Gerasimos Lampouras, Milan Gritta et al.

We present PanGu-Coder, a pretrained decoder-only language model adopting the PanGu-Alpha architecture for text-to-code generation, i.e. the synthesis of programming language solutions given a natural language problem description. We train PanGu-Coder using a two-stage strategy: the first stage employs Causal Language Modelling (CLM) to pre-train on raw programming language data, while the second stage uses a combination of Causal Language Modelling and Masked Language Modelling (MLM) training objectives that focus on the downstream task of text-to-code generation and train on loosely curated pairs of natural language program definitions and code functions. Finally, we discuss PanGu-Coder-FT, which is fine-tuned on a combination of competitive programming problems and code with continuous integration tests. We evaluate PanGu-Coder with a focus on whether it generates functionally correct programs and demonstrate that it achieves equivalent or better performance than similarly sized models, such as CodeX, while attending a smaller context window and training on less data.

Songbo Hu, Han Zhou, Moy Yuan et al. · cambridge

Achieving robust language technologies that can perform well across the world's many languages is a central goal of multilingual NLP. In this work, we take stock of and empirically analyse task performance disparities that exist between multilingual task-oriented dialogue (ToD) systems. We first define new quantitative measures of absolute and relative equivalence in system performance, capturing disparities across languages and within individual languages. Through a series of controlled experiments, we demonstrate that performance disparities depend on a number of factors: the nature of the ToD task at hand, the underlying pretrained language model, the target language, and the amount of ToD annotated data. We empirically prove the existence of the adaptation and intrinsic biases in current ToD systems: e.g., ToD systems trained for Arabic or Turkish using annotated ToD data fully parallel to English ToD data still exhibit diminished ToD task performance. Beyond providing a series of insights into the performance disparities of ToD systems in different languages, our analyses offer practical tips on how to approach ToD data collection and system development for new languages.

Milan Gritta, Ruoyu Hu, Ignacio Iacobacci

Task-oriented personal assistants enable people to interact with a host of devices and services using natural language. One of the challenges of making neural dialogue systems available to more users is the lack of training data for all but a few languages. Zero-shot methods try to solve this issue by acquiring task knowledge in a high-resource language such as English with the aim of transferring it to the low-resource language(s). To this end, we introduce CrossAligner, the principal method of a variety of effective approaches for zero-shot cross-lingual transfer based on learning alignment from unlabelled parallel data. We present a quantitative analysis of individual methods as well as their weighted combinations, several of which exceed state-of-the-art (SOTA) scores as evaluated across nine languages, fifteen test sets and three benchmark multilingual datasets. A detailed qualitative error analysis of the best methods shows that our fine-tuned language models can zero-shot transfer the task knowledge better than anticipated.

Debjit Paul, Daniel Murphy, Milan Gritta et al.

Large language model (LLM)-based agents are increasingly used to solve complex tasks involving tool use, such as web browsing, code execution, and data analysis. However, current evaluation benchmarks do not adequately assess their ability to solve real-world tasks that require synthesizing information from multiple sources and inferring insights beyond simple fact retrieval. To address this, we introduce DEEPSYNTH, a novel benchmark designed to evaluate agents on realistic, time-consuming problems that combine information gathering, synthesis, and structured reasoning to produce insights. DEEPSYNTH contains 120 tasks collected across 7 domains and data sources covering 67 countries. DEEPSYNTH is constructed using a multi-stage data collection pipeline that requires annotators to collect official data sources, create hypotheses, perform manual analysis, and design tasks with verifiable answers. When evaluated on DEEPSYNTH, 11 state-of-the-art LLMs and deep research agents achieve a maximum F1 score of 8.97 and 17.5 on the LLM-judge metric, underscoring the difficulty of the benchmark. Our analysis reveals that current agents struggle with hallucinations and reasoning over large information spaces, highlighting DEEPSYNTH as a crucial benchmark for guiding future research.

Leonidas Gee, Milan Gritta, Gerasimos Lampouras et al.

Code Language Models have been trained to generate accurate solutions, typically with no regard for runtime. On the other hand, previous works that explored execution optimisation have observed corresponding drops in functional correctness. To that end, we introduce Code-Optimise, a framework that incorporates both correctness (passed, failed) and runtime (quick, slow) as learning signals via self-generated preference data. Our framework is both lightweight and robust as it dynamically selects solutions to reduce overfitting while avoiding a reliance on larger models for learning signals. Code-Optimise achieves significant improvements in pass@k while decreasing the competitive baseline runtimes by an additional 6% for in-domain data and up to 3% for out-of-domain data. As a by-product, the average length of the generated solutions is reduced by up to 48% on MBPP and 23% on HumanEval, resulting in faster and cheaper inference. The generated data and codebase is open-sourced at https://github.com/huawei-noah/HEBO/tree/Code_Optimise.

Milan Gritta, Gerasimos Lampouras, Ignacio Iacobacci

Language models (LMs) as conversational assistants recently became popular tools that help people accomplish a variety of tasks. These typically result from adapting LMs pretrained on general domain text sequences through further instruction-tuning and possibly preference optimisation methods. The evaluation of such LMs would ideally be performed using human judgement, however, this is not scalable. On the other hand, automatic evaluation featuring auxiliary LMs as judges and/or knowledge-based tasks is scalable but struggles with assessing conversational ability and adherence to instructions. To help accelerate the development of LMs as conversational assistants, we propose a novel automatic evaluation task: HumanRankEval (HRE). It consists of a large-scale, diverse and high-quality set of questions, each with several answers authored and scored by humans. To perform evaluation, HRE ranks these answers based on their log-likelihood under the LM's distribution, and subsequently calculates their correlation with the corresponding human rankings. We support HRE's efficacy by investigating how efficiently it separates pretrained and instruction-tuned LMs of various sizes. We show that HRE correlates well with human judgements and is particularly responsive to model changes following instruction-tuning.

Milan Gritta, Huiyin Xue, Gerasimos Lampouras

Speculative decoding (SD) accelerates Large Language Model (LLM) generation by using an efficient draft model to propose the next few tokens, which are verified by the LLM in a single forward call, reducing latency while preserving its outputs. We focus on retrieval-based SD where the draft model retrieves the next tokens from a non-parametric datastore. Sparse retrieval (REST), which operates on the surface form of strings, is currently the dominant paradigm due to its simplicity and scalability. However, its effectiveness is limited due to the usage of short contexts and exact string matching. Instead, we introduce Dense Retrieval for Speculative Decoding (DReSD), a novel framework that uses approximate nearest neighbour search with contextualised token embeddings to retrieve the most semantically relevant token sequences for SD. Extensive experiments show that DReSD achieves (on average) 87% higher acceptance rates, 65% longer accepted tokens and 19% faster generation speeds compared to sparse retrieval (REST).

Meiru Zhang, Philipp Borchert, Milan Gritta et al.

Automating the formalization of mathematical statements for theorem proving remains a major challenge for Large Language Models (LLMs). LLMs struggle to identify and utilize the prerequisite mathematical knowledge and its corresponding formal representation in languages like Lean. Current retrieval-augmented autoformalization methods query external libraries using the informal statement directly, but overlook a fundamental limitation: informal mathematical statements are often complex and offer limited context on the underlying math concepts. To address this, we introduce DRIFT, a novel framework that enables LLMs to decompose informal mathematical statements into smaller, more tractable ''sub-components''. This facilitates targeted retrieval of premises from mathematical libraries such as Mathlib. Additionally, DRIFT retrieves illustrative theorems to help models use premises more effectively in formalization tasks. We evaluate DRIFT across diverse benchmarks (ProofNet, ConNF, and MiniF2F-test) and find that it consistently improves premise retrieval, nearly doubling the F1 score compared to the DPR baseline on ProofNet. Notably, DRIFT demonstrates strong performance on the out-of-distribution ConNF benchmark, with BEq+@10 improvements of 37.14% and 42.25% using GPT-4.1 and DeepSeek-V3.1, respectively. Our analysis shows that retrieval effectiveness in mathematical autoformalization depends heavily on model-specific knowledge boundaries, highlighting the need for adaptive retrieval strategies aligned with each model's capabilities.

Benjamin Minixhofer, Milan Gritta, Ignacio Iacobacci

Transfer learning has become the dominant paradigm for many natural language processing tasks. In addition to models being pretrained on large datasets, they can be further trained on intermediate (supervised) tasks that are similar to the target task. For small Natural Language Inference (NLI) datasets, language modelling is typically followed by pretraining on a large (labelled) NLI dataset before fine-tuning with each NLI subtask. In this work, we explore Gradient Boosted Decision Trees (GBDTs) as an alternative to the commonly used Multi-Layer Perceptron (MLP) classification head. GBDTs have desirable properties such as good performance on dense, numerical features and are effective where the ratio of the number of samples w.r.t the number of features is low. We then introduce FreeGBDT, a method of fitting a GBDT head on the features computed during fine-tuning to increase performance without additional computation by the neural network. We demonstrate the effectiveness of our method on several NLI datasets using a strong baseline model (RoBERTa-large with MNLI pretraining). The FreeGBDT shows a consistent improvement over the MLP classification head.

Milan Gritta, Ignacio Iacobacci

The introduction of pretrained cross-lingual language models brought decisive improvements to multilingual NLP tasks. However, the lack of labelled task data necessitates a variety of methods aiming to close the gap to high-resource languages. Zero-shot methods in particular, often use translated task data as a training signal to bridge the performance gap between the source and target language(s). We introduce XeroAlign, a simple method for task-specific alignment of cross-lingual pretrained transformers such as XLM-R. XeroAlign uses translated task data to encourage the model to generate similar sentence embeddings for different languages. The XeroAligned XLM-R, called XLM-RA, shows strong improvements over the baseline models to achieve state-of-the-art zero-shot results on three multilingual natural language understanding tasks. XLM-RA's text classification accuracy exceeds that of XLM-R trained with labelled data and performs on par with state-of-the-art models on a cross-lingual adversarial paraphrasing task.

Milan Gritta, Gerasimos Lampouras, Ignacio Iacobacci

Task-oriented dialogue systems typically rely on large amounts of high-quality training data or require complex handcrafted rules. However, existing datasets are often limited in size considering the complexity of the dialogues. Additionally, conventional training signal inference is not suitable for non-deterministic agent behaviour, i.e. considering multiple actions as valid in identical dialogue states. We propose the Conversation Graph (ConvGraph), a graph-based representation of dialogues that can be exploited for data augmentation, multi-reference training and evaluation of non-deterministic agents. ConvGraph generates novel dialogue paths to augment data volume and diversity. Intrinsic and extrinsic evaluation across three datasets shows that data augmentation and/or multi-reference training with ConvGraph can improve dialogue success rates by up to 6.4%.

Milan Gritta

We undertake the task of comparing lexicon-based sentiment classification of film reviews with machine learning approaches. We look at existing methodologies and attempt to emulate and improve on them using a 'given' lexicon and a bag-of-words approach. We also utilise syntactical information such as part-of-speech and dependency relations. We will show that a simple lexicon-based classification achieves good results however machine learning techniques prove to be the superior tool. We also show that more features do not necessarily deliver better performance as well as elaborate on three further enhancements not tested in this article.

Milan Gritta, Mohammad Taher Pilehvar, Nigel Collier

Empirical methods in geoparsing have thus far lacked a standard evaluation framework describing the task, metrics and data used to compare state-of-the-art systems. Evaluation is further made inconsistent, even unrepresentative of real-world usage by the lack of distinction between the different types of toponyms, which necessitates new guidelines, a consolidation of metrics and a detailed toponym taxonomy with implications for Named Entity Recognition (NER) and beyond. To address these deficiencies, our manuscript introduces a new framework in three parts. Part 1) Task Definition: clarified via corpus linguistic analysis proposing a fine-grained Pragmatic Taxonomy of Toponyms. Part 2) Metrics: discussed and reviewed for a rigorous evaluation including recommendations for NER/Geoparsing practitioners. Part 3) Evaluation Data: shared via a new dataset called GeoWebNews to provide test/train examples and enable immediate use of our contributions. In addition to fine-grained Geotagging and Toponym Resolution (Geocoding), this dataset is also suitable for prototyping and evaluating machine learning NLP models.