Interview with 2025 ICE Award winners for Structures and Buildings
We are delighted to feature an interview with the winners of the 2025 ICE Award Frederick Palmer Prize, which recognises the best paper published in Structures and Buildings during 2024. This recognition celebrates outstanding academic research that advances the field of structural engineering. In this brief conversation, we explore the technical contributions of the paper as well as highlight the broader impact of the research and the journal. The authors share insights into their award-winning work, the motivation behind it, and their experience publishing with ICE Publishing/Emerald.
From Left: Nima Khodadadi, Siamak Talatahari, Amir H Gandomi
Professor Amir H Gandomi, Dr Siamak Talatahari, Dr Nima Khodadadi Congratulations on your recent award! You are the recipients of 2024 Structures and Buildings Best Paper for article ‘ANNA: advanced neural network algorithm for optimization of structures’. Could you briefly introduce yourself and your background to our readers?
First, we would like to sincerely thank ICE Publishing and Emerald for this prestigious recognition. It is a great honor to receive the 2024 Structures and Buildings Best Paper Award. This acknowledgment reflects our team's shared commitment to advancing the frontiers of structural engineering through interdisciplinary research and innovation. Our team brings together a diverse and complementary set of expertise at the intersection of structural engineering, optimization, and artificial intelligence.
Professor Amir H. Gandomi, based at the University of Technology Sydney (UTS), is internationally recognized for his contributions to machine learning, evolutionary computation, and their applications in engineering.
Dr. Siamak TalatAhari, affiliated with Macquarie University, has a background in both structural engineering and data science, with particular strengths in optimization techniques and AI-integrated modeling for engineering systems.
Dr. Nima Khodadadi, currently a postdoctoral researcher at the University of California, Berkeley, brings a strong background in computational mechanics and intelligent modeling of complex structural systems.
Collectively, our research focuses on the development of advanced computational tools aimed at improving the performance, resilience, and sustainability of civil infrastructure. We are particularly interested in integrating data-driven techniques, including deep learning, surrogate modeling, and metaheuristic optimization, into various phases of structural design and assessment.
Our work lies at the convergence of artificial intelligence and structural engineering, with a focus on creating adaptive and intelligent frameworks for analyzing and optimizing structures. By embedding AI capabilities into traditional engineering workflows, we aim to tackle challenges such as uncertainty, nonlinearity, and computational cost more effectively.
This recognition is not only a validation of our past work but also a motivation to continue pushing the boundaries of what is possible when engineering expertise is combined with emerging technologies. Our long-term vision is to help shape the future of civil infrastructure through intelligent, efficient, and sustainable design strategies.
What inspired the development of ANNA, and how does it differ from existing neural network approaches in structural? What key question were you aiming to answer with your research?
The development of ANNA was motivated by the increasing demand for efficient, flexible, and scalable optimization tools capable of handling complex engineering problems. While traditional methods have long been applied in the field, ANNA stands out as a robust optimization algorithm that, although inspired by neural behavior, is fundamentally rooted in metaheuristic search principles. Our core objective was to tackle a persistent challenge in optimization: how to effectively balance global exploration and local exploitation within high-dimensional, nonlinear, and multi-modal design spaces. Many existing algorithms require numerous iterations and often struggle with generalization, especially in complex or constrained scenarios. In contrast, ANNA offers a simple yet powerful framework that operates directly on structural performance functions, eliminating the need for extensive training data.
What sets ANNA apart is its adaptive nature which optimization agents are guided by dynamic learning mechanisms that emulate the self-adjusting behavior of biological neurons. This adaptive strategy allows ANNA to regulate its exploration and exploitation tendencies over time, leading to faster convergence and greater solution diversity compared to many conventional metaheuristics.
The key question we set out to explore was: 'Can we develop an intelligent, learning-inspired optimizer that performs robustly across a wide range of structural optimization problems, without relying on problem-specific tuning or large training datasets?'
Our results demonstrate that ANNA not only achieves competitive or superior performance compared to conventional methods, but also offers enhanced adaptability, making it a promising tool for structural design applications where efficiency, robustness, and generality are critical.
Could you walk us through the most significant challenges you faced during the development and validation of ANNA?
One of the main challenges was designing a learning-inspired mechanism that mimics neural behavior while remaining simple and effective for optimization. We had to ensure ANNA could adapt dynamically without relying on large datasets or complex training. Another challenge was ensuring robustness and scalability across diverse and complex problems. This required extensive testing to maintain stability and performance under various conditions. Finally, validating ANNA involved rigorous benchmarking and statistical comparisons with established algorithms, while also keeping the method lightweight and easy to implement. Balancing adaptability, simplicity, and effectiveness was central throughout the development process.
Your paper demonstrates accuracy in damage detection. How scalable is ANNA for real-world infrastructure monitoring, especially in large-scale or aging structures?
While our paper does not focus directly on damage detection, ANNA was developed as a general-purpose optimization method. In this study, it was applied to structural design with the goal of achieving safe, efficient, and cost-effective solutions, ultimately helping to prevent damage through improved design. Its strength lies in optimizing complex, high-dimensional engineering problems with minimal assumptions. That said, ANNA is a generalizable framework and can be extended to related fields such as health monitoring and damage detection. Its adaptive nature and ability to handle nonlinear, multi-objective problems make it suitable for integration with data-driven models or sensor-based systems in infrastructure monitoring. In terms of scalability, ANNA has been tested on large-scale structural design problems and has shown strong performance and stability. With further customization, it holds potential for real-world applications in aging infrastructure, particularly when coupled with real-time data or surrogate models.
How do you envision ANNA evolving in the next few years? Are there plans to integrate it with other technologies like digital twins or IoT-based monitoring systems?
We see strong potential for ANNA to evolve as a core optimization engine within next-generation intelligent engineering systems. Thanks to its adaptability and minimal reliance on predefined parameters, ANNA is well-suited for integration with emerging technologies such as digital twins and IoT-based structural monitoring platforms. In the context of digital twins, ANNA can be used to continuously optimize structural designs or operational strategies in real time, as new data flows in from sensors or simulations. Similarly, when combined with IoT-enabled infrastructure, ANNA could support adaptive decision-making for maintenance, load management, or early-warning systems, closing the loop between design, monitoring, and action. Future developments may also include hybrid versions of ANNA that incorporate surrogate modeling, physics-informed learning, or multi-fidelity data, further enhancing its efficiency in real-world applications. Ultimately, our vision is to make ANNA not just a standalone optimizer but a flexible, intelligent module in the digital transformation of infrastructure systems.
Winning the Best Paper of the Year is a major achievement: what does this recognition mean to you and your research team? Are there any emerging areas or interdisciplinary connections you are excited to explore?
We are truly honored to receive the Best Paper of the Year award. This recognition is a meaningful validation of our efforts to push the boundaries of engineering optimization through innovative and interdisciplinary approaches. It reflects the hard work, collaboration, and shared vision of our team, and it motivates us to continue advancing research that bridges traditional structural engineering with modern AI-driven methodologies. Looking ahead, we are particularly excited about exploring interdisciplinary connections between optimization, digital twin technology, generative AI, and real-time monitoring systems. The convergence of these fields presents powerful opportunities to create intelligent, adaptive infrastructure systems. We’re also interested in applying ANNA to broader domains, including resilient infrastructure planning, energy-efficient design, and AI-enhanced decision-making in engineering practice. This award encourages us to keep innovating at the intersection of engineering, data science, and artificial intelligence, and to contribute to shaping the future of smart and sustainable infrastructure.
Why did you choose Structures and Buildings as the journal for this work, and how do you see the journal’s role in advancing innovation in structural engineering?
We chose Structures and Buildings because of its strong reputation as a leading journal that bridges cutting-edge structural engineering research with practical application. Its commitment to publishing high-quality, impactful work makes it an ideal platform for innovations like ANNA, which integrate computational intelligence into structural design and optimization. The journal’s broad readership and interdisciplinary focus support the convergence of traditional engineering methods with emerging technologies such as AI, digital tools, and smart infrastructure systems. By fostering a forward-looking community and emphasizing industry relevance, Structures and Buildings plays a vital role in advancing innovation and shaping the future of structural engineering.
What advice would you give to early-career researchers working at the intersection of AI and civil engineering?
Our advice would be to embrace both depth and breadth: build a strong foundation in engineering principles while actively learning the core concepts of AI, data science, and optimization. The true impact lies in understanding both domains well enough to bridge them effectively. Focus on real-world problems, seek collaborations with industry or applied research projects where AI can offer meaningful value. Be patient with interdisciplinary challenges; integration takes time and often requires translating between very different ways of thinking. Finally, stay curious and open to new tools, but don’t adopt AI just for the sake of novelty. Aim for solutions that are interpretable, reliable, and practical, as these will have the greatest impact in engineering practice.
******
Professor Amir H Gandomi, Dr Siamak Talatahari, Dr Nima Khodadadi‘s award winning article, 'ANNA: advanced neural network algorithm for optimization of structures' published in Volume 177, Issue 6 (June 2024) of Structures and Buildings, will be free to read for a year. All Structures and Buildings issues are available on Emerald Insight .
For more awards on related engineering subjects, please visit ICE Publishing Awards.