Luke Dickens

Jeya Vikranth Jeyakumar, Luke Dickens, Luis Garcia et al.

Recent efforts in interpretable deep learning models have shown that concept-based explanation methods achieve competitive accuracy with standard end-to-end models and enable reasoning and intervention about extracted high-level visual concepts from images, e.g., identifying the wing color and beak length for bird-species classification. However, these concept bottleneck models rely on a necessary and sufficient set of predefined concepts-which is intractable for complex tasks such as video classification. For complex tasks, the labels and the relationship between visual elements span many frames, e.g., identifying a bird flying or catching prey-necessitating concepts with various levels of abstraction. To this end, we present CoDEx, an automatic Concept Discovery and Extraction module that rigorously composes a necessary and sufficient set of concept abstractions for concept-based video classification. CoDEx identifies a rich set of complex concept abstractions from natural language explanations of videos-obviating the need to predefine the amorphous set of concepts. To demonstrate our method's viability, we construct two new public datasets that combine existing complex video classification datasets with short, crowd-sourced natural language explanations for their labels. Our method elicits inherent complex concept abstractions in natural language to generalize concept-bottleneck methods to complex tasks.

Binxia Xu, Antonis Bikakis, Daniel Onah et al.

Given that AI systems are set to play a pivotal role in future decision-making processes, their trustworthiness and reliability are of critical concern. Due to their scale and complexity, modern AI systems resist direct interpretation, and alternative ways are needed to establish trust in those systems, and determine how well they align with human values. We argue that good measures of the information processing similarities between AI and humans, may be able to achieve these same ends. While Representational alignment (RA) approaches measure similarity between the internal states of two systems, the associated data can be expensive and difficult to collect for human systems. In contrast, Behavioural alignment (BA) comparisons are cheaper and easier, but questions remain as to their sensitivity and reliability. We propose two new behavioural alignment metrics misclassification agreement which measures the similarity between the errors of two systems on the same instances, and class-level error similarity which measures the similarity between the error distributions of two systems. We show that our metrics correlate well with RA metrics, and provide complementary information to another BA metric, within a range of domains, and set the scene for a new approach to value alignment.

Antonis Bikakis, Aissatou Diallo, Luke Dickens et al.

Whilst cooking is a very important human activity, there has been little consideration given to how we can formalize recipes for use in a reasoning framework. We address this need by proposing a graphical formalization that captures the comestibles (ingredients, intermediate food items, and final products), and the actions on comestibles in the form of a labelled bipartite graph. We then propose formal definitions for comparing recipes, for composing recipes from subrecipes, and for deconstructing recipes into subrecipes. We also introduce and compare two formal definitions for substitution into recipes which are required when there are missing ingredients, or some actions are not possible, or because there is a need to change the final product somehow.

Aissatou Diallo, Antonis Bikakis, Luke Dickens et al.

Understanding procedural texts, such as cooking recipes, is essential for enabling machines to follow instructions and reason about tasks, a key aspect of intelligent reasoning. In cooking, these instructions can be interpreted as a series of modifications to a food preparation. For a model to effectively reason about cooking recipes, it must accurately discern and understand the inputs and outputs of intermediate steps within the recipe. We present a new corpus of cooking recipes enriched with descriptions of intermediate steps that describe the input and output for each step. PizzaCommonsense serves as a benchmark for the reasoning capabilities of LLMs because it demands rigorous explicit input-output descriptions to demonstrate the acquisition of implicit commonsense knowledge, which is unlikely to be easily memorized. GPT-4 achieves only 26\% human-evaluated preference for generations, leaving room for future improvements.

Aissatou Diallo, Antonis Bikakis, Luke Dickens et al.

In this paper, we introduce Rule-Guided Feedback (RGF), a framework designed to enhance Large Language Model (LLM) performance through structured rule adherence and strategic information seeking. RGF implements a teacher-student paradigm where rule-following is forced through established guidelines. Our framework employs a Teacher model that rigorously evaluates each student output against task-specific rules, providing constructive guidance rather than direct answers when detecting deviations. This iterative feedback loop serves two crucial purposes: maintaining solutions within defined constraints and encouraging proactive information seeking to resolve uncertainties. We evaluate RGF on diverse tasks including Checkmate-in-One puzzles, Sonnet Writing, Penguins-In-a-Table classification, GSM8k, and StrategyQA. Our findings suggest that structured feedback mechanisms can significantly enhance LLMs' performance across various domains.

Aissatou Diallo, Antonis Bikakis, Luke Dickens et al.

An interesting class of commonsense reasoning problems arises when people are faced with natural disasters. To investigate this topic, we present \textsf{RESPONSE}, a human-curated dataset containing 1789 annotated instances featuring 6037 sets of questions designed to assess LLMs' commonsense reasoning in disaster situations across different time frames. The dataset includes problem descriptions, missing resources, time-sensitive solutions, and their justifications, with a subset validated by environmental engineers. Through both automatic metrics and human evaluation, we compare LLM-generated recommendations against human responses. Our findings show that even state-of-the-art models like GPT-4 achieve only 37\% human-evaluated correctness for immediate response actions, highlighting significant room for improvement in LLMs' ability for commonsense reasoning in crises.

Aissatou Diallo, Antonis Bikakis, Luke Dickens et al.

While Large Language Models (LLMs) demonstrate impressive reasoning capabilities, understanding and validating their knowledge utilization remains challenging. Chain-of-thought (CoT) prompting partially addresses this by revealing intermediate reasoning steps, but the knowledge flow and application remain implicit. We introduce IAO (Input-Action-Output) prompting, a structured template-based method that explicitly models how LLMs access and apply their knowledge during complex reasoning tasks. IAO decomposes problems into sequential steps, each clearly identifying the input knowledge being used, the action being performed, and the resulting output. This structured decomposition enables us to trace knowledge flow, verify factual consistency, and identify potential knowledge gaps or misapplications. Through experiments across diverse reasoning tasks, we demonstrate that IAO not only improves zero-shot performance but also provides transparency in how LLMs leverage their stored knowledge. Human evaluation confirms that this structured approach enhances our ability to verify knowledge utilization and detect potential hallucinations or reasoning errors. Our findings provide insights into both knowledge representation within LLMs and methods for more reliable knowledge application.

Aissatou Diallo, Antonis Bikakis, Luke Dickens et al.

Cooking recipes are one of the most readily available kinds of procedural text. They consist of natural language instructions that can be challenging to interpret. In this paper, we propose a model to identify relevant information from recipes and generate a graph to represent the sequence of actions in the recipe. In contrast with other approaches, we use an unsupervised approach. We iteratively learn the graph structure and the parameters of a $\mathsf{GNN}$ encoding the texts (text-to-graph) one sequence at a time while providing the supervision by decoding the graph into text (graph-to-text) and comparing the generated text to the input. We evaluate the approach by comparing the identified entities with annotated datasets, comparing the difference between the input and output texts, and comparing our generated graphs with those generated by state of the art methods.