Author name cluster

Mark Dras

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

7 papers

2 author rows

AAAI Conference 2026 System Paper

Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA

Yiran Zhang
Mingyang Lin
Mark Dras
Usman Naseem

Recent research has increasingly focused on the reasoning capabilities of Large Language Models (LLMs) in multi-turn interactions, as these scenarios more closely mirror real-world problem-solving. However, analyzing the intricate reasoning processes within these interactions presents a significant challenge due to complex contextual dependencies and a lack of specialized visualization tools, leading to a high cognitive load for researchers. To address this gap, we present VISTA, an web-based Visual Interactive System for Textual Analytics in multi-turn reasoning tasks. VISTA allows users to visualize the influence of context on model decisions and interactively modify conversation histories to conduct "what-if" analyses across different models. Furthermore, the platform can automatically parse a session and generate a reasoning dependency tree, offering a transparent view of the model's step-by-step logical path. By providing a unified and interactive framework, VISTA significantly reduces the complexity of analyzing reasoning chains, thereby facilitating a deeper understanding of the capabilities and limitations of current LLMs. The platform is open-source and supports easy integration of custom benchmarks and local models.

PDF Details DOI

ECAI Conference 2025 Conference Paper

Bi-Directional Model Cascading with Proxy Confidence

David Warren
Mark Dras

Model Cascading, recently applied successfully to LLMs, is a simple but powerful technique that improves the efficiency of inference by selectively applying models of varying sizes. Models are used in sequence from smallest to largest, only deferring samples to large, costly models when smaller models are not sufficiently confident. Existing approaches to deferral use only limited small model confidence estimates because of the inaccessibility of the large model, although large model confidence is known to be important. We therefore propose a bi-directional approach to deferral that considers the confidence of small and large models in the cascade simultaneously through the use of a proxy for the large model. This requires a richer representation of model confidence to enable comparative calibration: we use an analysis of hidden states to improve post-invocation confidence of the small model, which in itself improves cascading results over prior approaches. We then combine this with a tiny proxy model to estimate pre-invocation confidence of the large model. We examine the proposed cascading system over challenging, multiple-choice datasets, finding improvements over standard cascading baselines reflected in reductions in deferrals to more costly models.

Details

TMLR Journal 2024 Journal Article

Temporally Rich Deep Learning Models for Magnetoencephalography

Tim Chard
Mark Dras
Paul Sowman
Steve Cassidy
Jia Wu

Deep learning has been used in a wide range of applications, but it has only very recently been applied to Magnetoencephalography (MEG). MEG is a neurophysiological technique used to investigate a variety of cognitive processes such as language and learning, and an emerging technology in the quest to identify neural correlates of cognitive impairments such as those occurring in dementia. Recent work has shown that it is possible to apply deep learning to MEG to categorise induced responses to stimuli across subjects. While novel in the application of deep learning, such work has generally used relatively simple neural network (NN) models compared to those being used in domains such as computer vision and natural language processing. In these other domains, there is a long history in developing complex NN models that combine spatial and temporal information. We propose more complex NN models that focus on modelling temporal relationships in the data, and apply them to the challenges of MEG data. We apply these models to an extended range of MEG-based tasks, and find that they substantially outperform existing work on a range of tasks, particularly but not exclusively temporally-oriented ones. We also show that an autoencoder-based preprocessing component that focuses on the temporal aspect of the data can improve the performance of existing models. Our source code is available at https://github.com/tim-chard/DeepLearningForMEG.

PDF Details

IJCAI Conference 2023 Conference Paper

OptIForest: Optimal Isolation Forest for Anomaly Detection

Haolong Xiang
Xuyun Zhang
Hongsheng Hu
Lianyong Qi
Wanchun Dou
Mark Dras
Amin Beheshti
Xiaolong Xu

Anomaly detection plays an increasingly important role in various fields for critical tasks such as intrusion detection in cybersecurity, financial risk detection, and human health monitoring. A variety of anomaly detection methods have been proposed, and a category based on the isolation forest mechanism stands out due to its simplicity, effectiveness, and efficiency, e. g. , iForest is often employed as a state-of-the-art detector for real deployment. While the majority of isolation forests use the binary structure, a framework LSHiForest has demonstrated that the multi-fork isolation tree structure can lead to better detection performance. However, there is no theoretical work answering the fundamentally and practically important question on the optimal tree structure for an isolation forest with respect to the branching factor. In this paper, we establish a theory on isolation efficiency to answer the question and determine the optimal branching factor for an isolation tree. Based on the theoretical underpinning, we design a practical optimal isolation forest OptIForest incorporating clustering based learning to hash which enables more information to be learned from data for better isolation quality. The rationale of our approach relies on a better bias-variance trade-off achieved by bias reduction in OptIForest. Extensive experiments on a series of benchmarking datasets for comparative and ablation studies demonstrate that our approach can efficiently and robustly achieve better detection performance in general than the state-of-the-arts including the deep learning based methods.

PDF Details DOI

NeurIPS Conference 2021 Conference Paper

Neural Rule-Execution Tracking Machine For Transformer-Based Text Generation

Yufei Wang
Can Xu
Huang Hu
Chongyang Tao
Stephen Wan
Mark Dras
Mark Johnson
Daxin Jiang

Sequence-to-Sequence (Seq2Seq) neural text generation models, especially the pre-trained ones (e. g. , BART and T5), have exhibited compelling performance on various natural language generation tasks. However, the black-box nature of these models limits their application in tasks where specific rules (e. g. , controllable constraints, prior knowledge) need to be executed. Previous works either design specific model structures (e. g. , Copy Mechanism corresponding to the rule "the generated output should include certain words in the source input'') or implement specialized inference algorithms (e. g. , Constrained Beam Search) to execute particular rules through the text generation. These methods require the careful design case-by-case and are difficult to support multiple rules concurrently. In this paper, we propose a novel module named Neural Rule-Execution Tracking Machine (NRETM) that can be equipped into various transformer-based generators to leverage multiple rules simultaneously to guide the neural generation model for superior generation performance in an unified and scalable way. Extensive experiments on several benchmarks verify the effectiveness of our proposed model in both controllable and general text generation tasks.

PDF Details

JAIR Journal 2020 Journal Article

Image Captioning using Facial Expression and Attention

Omid Mohamad Nezami
Mark Dras
Stephen Wan
Cecile Paris

Benefiting from advances in machine vision and natural language processing techniques, current image captioning systems are able to generate detailed visual descriptions. For the most part, these descriptions represent an objective characterisation of the image, although some models do incorporate subjective aspects related to the observer’s view of the image, such as sentiment; current models, however, usually do not consider the emotional content of images during the caption generation process. This paper addresses this issue by proposing novel image captioning models which use facial expression features to generate image captions. The models generate image captions using long short-term memory networks applying facial features in addition to other visual features at different time steps. We compare a comprehensive collection of image captioning models with and without facial features using all standard evaluation metrics. The evaluation metrics indicate that applying facial features with an attention mechanism achieves the best performance, showing more expressive and more correlated image captions, on an image caption dataset extracted from the standard Flickr 30K dataset, consisting of around 11K images containing faces. An analysis of the generated captions finds that, perhaps unexpectedly, the improvement in caption quality appears to come not from the addition of adjectives linked to emotional aspects of the images, but from more variety in the actions described in the captions.

PDF Details DOI

AAMAS Conference 2010 Conference Paper

Deceptive Agents and Language

Mark Dras
Debbie Richards
Meredith Taylor
Mary Gardiner

The use of virtual agents in training requires them to haveseveral human-like characteristics; one of these is the ability to appear deceptive. We take work from the psychologyliterature on cues to deception, with a focus on language-related cues, and examine whether it is possible to use resources from the field of Language Technology to constructscenarios with agents showing cues to deception detectableby human judges, a task that has been shown in a text-onlycontext to be difficult. We show that this detection is infact possible in the context of virtual agents, and that thereare interesting results for individual cues, in particular fordialogue- versus lexical-level cues, and a 'placebo' effect.

PDF