International Journal of Semantic Computing

Transdisciplinary practices in artificial intelligence (AI) requires good methodological techniques to bring researchers together from diverse disciplines. Different fields and traditions each may have an impact on the greater AI ecosystem, however, to properly gauge this impact, and include diverse voices, research methodologies must evolve to adopt new methods and techniques. To understand the shifting demands for new skills, we conducted a transdisciplinary workshop, bringing together stakeholders from industry, government and academia, by engaging with scenarios in healthcare and emergency response, using future visioning to imagine new skill requirements across job roles. We found that future visioning was a useful methodology to help frame discussions in a manner conducive to eliciting qualitative evidence of different policy approaches. Beyond individual skills, we propose the Human-AI Synergy Matrix as a collaborative competency framework, which integrates technical, professional, and strategic skills with levels of engagement such as awareness, communication, synergy, governance, and transformation. The matrix aims to facilitate evaluation and understanding of skills and critical issues in human-AI collaboration within AI ecosystems, emphasizing the need to address broader challenges in an interconnected manner.

The world has a considerable portion of its population living in vulnerable conditions, with about 47% of people living in poverty and approximately 9.3% in extreme poverty. To understand and face this problem, it is essential to have access to updated data on poverty and trace the socioeconomic profile of the population. In this context, Censuses and demographic surveys play a crucial role. However, conducting a Census in certain regions faces significant challenges, both in terms of cost and logistics. It is a complex process that demands significant resources and an extensive data collection effort throughout the national territories. Given the difficulties of collecting data in the traditional way, passively collected data sources, such as satellite imagery, may be an alternative way to measure these results. Therefore, we use computational techniques to combine nighttime and daytime satellite imagery to predict socioeconomic indicators. Our results demonstrate the efficacy of combining these features in predicting the average income and demographic density of cities across Brazil. In particular, the use of neural networks for inferring socioeconomic indicators has been demonstrated highly effective. This technique holds the potential to enhance our comprehension of the socioeconomic landscape in Brazil and provide technical means for the analysis of other countries that face difficulties in conducting Censuses. Our method has convincingly demonstrated that satellite images can be used and applied for socioeconomic purposes.

Uncontrolled wildfires pose a significant threat, potentially causing extensive damage to biodiversity, soil quality and human resources. It’s crucial to swiftly detect and predict these wildfires to minimize their catastrophic consequences. To address this, our research introduces a wildfire prediction model that ranks cities based on risk leveraging multi-criteria decision-making (MCDM) to systematically assess conflicting factors in decision-making. This model integrates wildfire risks into a city’s resilience strategy, utilizing fuzzy set theory to manage imprecise data and uncertainties. As part of this approach, we compile a new dataset encompassing weather patterns, vegetation types, terrain features and population density across various Californian cities. Ultimately, the model assesses and ranks the wildfire risk for each city in California.

As the capabilities of machine learning have developed, more researchers and health care providers are beginning to consider applications for health informatics to improve health care outcomes and address issues of health equity. The Centers for Medicare and Medicaid Services considers 30-day readmission rates to the Emergency Department (ED) to be an “outcome of care” measure. Such measures show how well a hospital is doing in preventing complications, educating patients about their care needs, and helping patients make a smooth transition from the hospital to home or other care facilities. While certain readmissions are medically necessary, hospitals usually aim to decrease the rate of 30-day ED readmissions by decreasing the number of avoidable unplanned revisits. This work is an evidential study that demonstrates the positive impact of integrating Community Health Workers (CHWs) and Social Determinants of Health (SDoH) in decreasing the 30-day unplanned hospital ED readmissions at Sinai Chicago. Using data from the Sinai Urban Health Institute, we compare predicting the readmissions of patients with and without data pertaining to CHWs and SDoH, characterize the improvement in predictions, and discuss lessons learned in the process. We show that when CHWs engage with patients, the predictive accuracy of the classifier is higher by 13.0%–15.2%. Importantly, we show that the features (data characteristics) related to the CHWs are important to the classification, pointing to the importance of the program. We optimize the classifier for engaged patients and demonstrate the improvement in predictive capabilities of the classier using multiple metrics and sets of features. We use our results to make recommendations for improving patient care, discuss limitations and future work. Notably, our work points directly to the human connection between patients and CHWs as an important feature predictive of readmission rate.

This paper introduces a method to solve the 2D irregular packing problem using Deep Reinforcement Learning (Deep RL) for logistics. Our method employs a Q agent trained to predict the best placement within a container, maximizing available space. Unlike previous Deep RL algorithms, our method introduces a dense reward function at each packing step, providing immediate feedback and accelerating learning. To our knowledge, this is the first approach to use a dense reward to address the 2D irregular packing problem. Building on our earlier work, we improve the deep neural network by incorporating the Double Deep Q-Network (DDQN) framework to enhance our deep Q-learning approach, reducing overestimation biases and improving decision-making reliability. Simulation results show the method’s effectiveness in completing the online 2D irregular packing tasks, achieving promising volume efficiency and packed piece metrics. This research extends our initial findings, highlighting the practical importance of DDQN and dense reward in advancing 2D irregular packing problem-solving. These advancements not only broaden the applications of deep learning but also hold practical importance for real-world logistics challenges.

In the era of big data, companies are increasingly driven to amass vast amounts of data, particularly in process industries where advanced sensor technologies are prevalent. However, obtaining accurate labels or product information through quality inspections can be prohibitively expensive. Active learning emerges as a promising approach to optimize data sampling by prioritizing the most informative data points. Nevertheless, active learning strategies heavily rely on predictive models that are iteratively updated. Aligning with the principles of data-centric AI, this study highlights the detrimental effects of passively incorporating all available process variables into a predictive model for guiding data collection. Specifically, in real-time sampling strategies based on online active learning, the inclusion of irrelevant features significantly hampers the efficiency of the learning process.

Graph analysis in large integrated circuit (IC) designs is an essential tool for verifying design logic and timing via dynamic analysis (DA). IC designs resemble graphs with each logic gate as a vertex and the conductive connections between gates as edges. Using DA digital statistical correlations, graph condensation, and graph partitioning, it is possible to identify anomalies in high-entropy component centers (HECCs) and paths within an IC design. Identification of HECC aids in DA signal integrity analysis when comparing design iterations which effectively lowers the computational complexity of DA within large IC graphs. In this paper, a devised methodology termed IC layout subgraph component center identification (CCI) is described. For design deviation analysis in DA, CCI lowers design computational complexity by identifying design deviations within an IC’s subgraphs. The CCI logic function signal integrity is verified using semi-supervised learning consisting of an unsupervised autoencoder anomaly detector to first identify anomalous subgraphs followed by efficient subgraph iterative refinement to locate specific deviated logic signals within the subgraphs.

Artificial intelligence (AI) faces a trifecta of grand challenges: the Energy Wall, the Alignment Problem and the Leap from Narrow AI to AGI. Contemporary AI solutions consume unsustainable amounts of energy during model training and daily operations. Making things worse, the amount of computation required to train each new AI model has been doubling every 2 months since 2020, directly translating to unprecedented increases in energy consumption. The leap from AI to AGI requires multiple functional subsystems operating in a balanced manner, which requires a system architecture. However, the current approach to AI lacks system design; even though system characteristics play a key role in the human brain; from the way it processes information to how it makes decisions. In this paper, we posit that system design is the missing piece in overcoming current AI the grand challenges. We present a Systematic Approach to AGI (SAGI) that utilizes system design principles to overcome the energy wall and the alignment challenges. This paper asserts that artificial intelligence can be realized through a multiplicity of design-specific pathways, rather than a singular, overarching AGI architecture. AGI systems may exhibit diverse architectural configurations and capabilities, contingent upon their intended use cases. We argue that AI alignment, the most difficult among the grand challenges, is not attainable without a way to reflect the complexity of the human moral system and its subsystems in the AGI architectures. We claim that AGI approaches such as symbolicism, connectionism and others are not fundamental to AGI but emergent from the system design processes. Hence, we focus on employing system design principles as a guiding framework, rather than solely concentrating on a universal AGI architecture.

Transformer-based large language models are currently at the forefront of modern artificial intelligence. Their prominence followed from the seminal paper Attention is All You Need [1]. Vaswani and his colleagues suggested placing attention mechanisms within the encoder and decoder modules of autoencoders rather than using them to focus between these two modules. In this paper we present first the seminal insights of early AI that lead to deep learning. We then describe the mathematical tools necessary for understanding the current generation of LLMs and follow this with a brief description of the transformer architecture. We then provide examples of LLMs in action and conclude with some observations of their promise and problems.