Tool invocation refusal - correctly refusing calls when no appropriate tool exists or tools cannot solve the task - is critical for agentic IR systems. This challenge becomes acute in low-resource languages deployed on small models, where weak multilingual representations compound the problem. We demonstrate that conventional fine-tuning shows inconsistent behavior: while improving tool-calling capability, it can degrade refusal metrics (Gemma-Thai: irrelevance -4pp, live hallucination -21pp). We propose structured reasoning forms: a training-time approach that teaches models to generate explicit reasoning about tool-invocation decisions through lightweight key-value schemas. Rather than simply adding more refusal examples, our method trains models to decompose decisions into interpretable stages (task understanding, tool assessment, risk evaluation), internalizing the reasoning process for ''when NOT to call'' rather than just ''how to call.'' Moreover, standard format-compliance evaluation masks refusal failures, artificially inflating baseline scores by approximately 30 percentage points because it incorrectly counts malformed tool-call attempts as correct refusals. On two multilingual 1B-parameter models evaluated on newly created Vietnamese and Thai benchmarks (BFCL-VN and BFCL-TH), models trained to generate reasoning forms improve irrelevance detection and live hallucination by roughly +10-30pp (reaching 50-60% and 43-61% respectively) while maintaining tool-calling capability. Compared with free-form CoT reasoning, structured reasoning achieves stronger refusal performance with 1.77-1.91× fewer reasoning tokens. Perplexity analysis demonstrates reasoning fundamentally changes internal model behavior, reducing hallucination probability from 67-76% to 1.2-1.3%.

Deployment of the Internet of Things (IoT) in real industrial scenarios is fraught with unseen challenges and planning intricacies. The inside engineering and methodologies behind a successful Industrial IoT (IIoT) setup are seldom discussed. In this paper, we describe a systems approach to identify, integrate, and deploy the different IIoT modules in a bottom-up manner. The process that go from first identifying and labelling the various assets to be monitored, to their abstraction and integration for a given industrial scenario, forms the major crux of the work presented in this paper. We deployed several sensor nodes in the Offshore Renewable Energy Catapult site, located in Blyth, UK. Each sensor node monitored a specific asset and communicated via LoRaWAN to our local Data Hub. A simple query interface and visualisation dashboard allowed real-time data assessment and rapid asset monitoring. To make the proposed approach easy to comprehend and applicable to other scenarios, we extracted a minimal translational bottom-up flowchart. The systems approach described here offers a robust methodology to plan out a real-world IIoT deployment.

Public perceptions of Robotics and Artificial Intelligence (RAI) are important in the acceptance, uptake, government regulation and research funding of this technology. Recent research has shown that the public's understanding of RAI can be negative or inaccurate. We believe effective public engagement can help ensure that public opinion is better informed. In this paper, we describe our first iteration of a high throughput in-person public engagement activity. We describe the use of a light touch quiz-format survey instrument to integrate in-the-wild research participation into the engagement, allowing us to probe both the effectiveness of our engagement strategy, and public perceptions of the future roles of robots and humans working in dangerous settings, such as in the off-shore energy sector. We critique our methods and share interesting results into generational differences within the public's view of the future of Robotics and AI in hazardous environments. These findings include that older peoples' views about the future of robots in hazardous environments were not swayed by exposure to our exhibit, while the views of younger people were affected by our exhibit, leading us to consider carefully in future how to more effectively engage with and inform older people.

Traditional Evolutionary Robotics (ER) employs evolutionary techniques to search for a single monolithic controller which can aid a robot to learn a desired task. These techniques suffer from boot-strap and deception issues when the tasks are complex for a single controller to learn. Behaviour-decomposition techniques have been used to divide a task into multiple subtasks and evolve separate subcontrollers for each subtask. However, these subcontrollers and the associated subcontroller arbitrator(s) are all evolved off-line. A distributed, fully embodied and evolutionary version of such approaches will greatly aid online learning and help reduce the reality gap. In this paper, we propose an immunology-inspired embodied action-evolution cum selection algorithm that can cater to distributed ER. This algorithm evolves different subcontrollers for different portions of the search space in a distributed manner just as antibodies are evolved and primed for different antigens in the antigenic space. Experimentation on a collective of real robots embodied with the algorithm showed that a repertoire of antibody-like subcontrollers was created, evolved and shared on-the-fly to cope up with different environmental conditions. In addition, instead of the conventionally used approach of broadcasting for sharing, we present an Intelligent Packet Migration scheme that reduces energy consumption.

Transfer Learning (TL) plays a crucial role when a given dataset has insufficient labeled examples to train an accurate model. In such scenarios, the knowledge accumulated within a model pre-trained on a source dataset can be transferred to a target dataset, resulting in the improvement of the target model. Though TL is found to be successful in the realm of image-based applications, its impact and practical use in Natural Language Processing (NLP) applications is still a subject of research. Due to their hierarchical architecture, Deep Neural Networks (DNN) provide flexibility and customization in adjusting their parameters and depth of layers, thereby forming an apt area for exploiting the use of TL. In this paper, we report the results and conclusions obtained from extensive empirical experiments using a Convolutional Neural Network (CNN) and try to uncover thumb rules to ensure a successful positive transfer. In addition, we also highlight the flawed means that could lead to a negative transfer. We explore the transferability of various layers and describe the effect of varying hyper-parameters on the transfer performance. Also, we present a comparison of accuracy value and model size against state-of-the-art methods. Finally, we derive inferences from the empirical results and provide best practices to achieve a successful positive transfer.

With robots entering the world of Cyber Physical Systems (CPS), ordering the execution of allocated tasks during runtime becomes crucial. This is so because, in the real world, there can be several physical tasks that use shared resources that need to be executed concurrently. In this article, we propose a mechanism to solve this issue of ordering task executions within a CPS that inherently handles mutual exclusion. The mechanism caters to a decentralized and distributed CPS comprising nodes such as computers, robots, and sensor nodes and uses mobile software agents that knit through them to aid the execution of the various tasks while also ensuring mutual exclusion of shared resources. The computations, communications, and control are achieved through these mobile agents. Physical execution of the tasks is performed by the robots in an asynchronous and pipelined manner without the use of a clock. The mechanism also features addition and deletion of tasks and insertion and removal of robots facilitating On-The-Fly Programming . As an application, a Warehouse Management System as a CPS has been implemented. The article concludes with the results and discussions on using the mechanism in both emulated and real-world environments.

Paradigms derived from the Biological Immune System have shown great promise in devising algorithms that can be used in a variety of fields. One such is the Idiotypic Network (IN) proposed by Jerne, wherein the antibodies form a virtual network based on both antigenic and mutual stimulations and suppressions which in turn change their respective concentrations. This paper portrays how INs can be augmented with information sharing and used within a Multi-Robot System (MRS), to learn the right behaviours and actions to achieve a goal inherent to an environment. Each robot in the MRS scenario is equipped with its own IN that shares the information learned from the environment with that of its peers. We also present a case when the environment changes during run-time and the MRS re-learns by making re-adjustments in the concentrations of a new set of antibodies. The results obtained from a real scenario comprising three robots using an IN each, portrayed herein, justify the viability of such an approach.