Machine learning, AI platforms, and biases

Understanding the nature of chemical bonding and its variation in strength across physically tunable factors is important for the development of novel catalytic materials. One way to speed up this process is to employ machine learning (ML) algorithms with online data repositories curated from high-throughput experiments or quantum-chemical simulations. Despite the reasonable predictive performance of ML models for predicting reactivity properties of solid surfaces, the ever-growing complexity of modern algorithms, e.g., deep learning, makes them black boxes with little to no explanation. In this Perspective, we discuss recent advances of interpretable ML for opening up these black boxes from the standpoints of feature engineering, algorithm development, and post hoc analysis. We underline the pivotal role of interpretability as the foundation of next-generation ML algorithms and emerging AI platforms for driving discoveries across scientific disciplines.

Edge computing enables lower latency and secure real-time processing for machine learning (ML) workloads. State-of-the-art edge platforms are heterogeneous systems with compute cores of diverse performance and energy capabilities. This diversity motivates the need to identify the strengths of individual compute cores for more informed heterogeneous implementations of workloads. In this study, we benchmark a state-of-the-art heterogeneous platform consisting of a multicore CPU, integrated GPU and NPU, on its performance on various ML workloads. We measure system performance in terms of latency, throughput, and Energy-Delay Product (EDP) to analyze the performance-energy tradeoffs of such systems across various configurations -independently on the CPU, GPU or NPU and a heterogeneous NPU-GPU configuration. Our results indicate the superiority of NPUs for energy-efficient inference of LLMs and CNNs. We find that the NPU offers up to 5.2× speedup over CPU for CNN inference. For LLM inference, the NPU achieves up to 4× speedup during the prefill stage, while the GPU offers a 22× speedup on the decode stage. We observe that hybrid NPU-GPU implementation of LLM inference is superior to CPU-only inference in terms of energy consumption and end-to-end (E2E) latency, with an EDP decrease of about 93%. We highlight the broader applicability of our results by discussing extensions to long-context inference in causal LLMs and emerging state-space models on heterogeneous platforms.

The integration of artificial intelligence with customer relationship management platforms in healthcare represents a transformative advancement in personalized medicine. This convergence enables healthcare providers to consolidate patient data from diverse sources into unified platforms, facilitating individualized diagnosis and treatment planning. Cloud-based technologies serve as essential infrastructure, providing the scalability and computational capacity necessary for managing complex healthcare data ecosystems. The implementation of standardized APIs and event-driven architectures supports real-time data synchronization across clinical systems, while machine learning algorithms identify subtle patterns within patient data that might otherwise remain undetected. Workflow automation enhances administrative efficiency, reducing documentation burden and enabling clinical staff to focus on direct patient care. The resulting framework supports a multidimensional approach to treatment personalization that incorporates not only clinical data but also patient preferences and social determinants of health, ultimately transforming healthcare delivery toward more responsive, patient-centered models.

As globalization progresses at a rapid pace, the dramatic advancements in AI technology have significantly accelerated its integration into global education systems. This study explores the interplay between AI and international education, analyzes the educational transformations driven by AI adoption. We examine the integration of AI technologies—such as natural language processing (NLP), machine learning, and generative AI—in addressing language barriers, personalizing learning experiences, optimizing teaching management, and fostering global educational collaboration. A systematic technical implementation framework is presented, encompassing local content processing, AI-generated content (AIGC) via cloud computing, and efficient multi-platform distribution. Case studies of leading AI platforms (e.g., ChatGPT, Grok, Qwen) demonstrate their capacity to enhance accessibility and engagement. Critical risks—including privacy vulnerabilities, algorithmic bias, cultural conflicts, and ethical concerns—are rigorously assessed, with mitigation strategies proposed. Our findings reveal that AI-driven international education not only enhances teaching efficiency and cost-effectiveness but also promotes equitable access and cross-cultural awareness. The research provides theoretical guidance for the future development of international education.

Deployed AI platforms typically ship with bulky system architectures which present bottlenecks and a high risk of failure. A serverless deployment can mitigate these factors and provide a cost-effective, automatically scalable (up or down) and elastic real-time on-demand AI solution. However, deploying high complexity production workloads into serverless environments is far from trivial, e.g., due to factors such as minimal allowance for physical codebase size, low amount of runtime memory, lack of GPU support and a maximum runtime before termination via timeout. In this paper we propose a set of optimization techniques and show how these transform a codebase which was previously incompatible with a serverless deployment into one that can be successfully deployed in a serverless environment; without compromising capability or performance. The techniques are illustrated via worked examples that have been deployed live on rail data and realtime predictions on train movements on the UK rail network. The similarities of a serverless environment to other resource constrained environments (IoT, Mobile) means the techniques can be applied to a range of use cases.

Reproducibility in robotics research requires capable, shared hardware platforms which can be used for a wide variety of research. We’ve seen the power of these sorts of shared platforms in more general machine learning research, where there is constant iteration on shared AI platforms like PyTorch. To be able to make rapid progress in robotics in the same way, we propose that we need: (1) shared real-world platforms which allow different teams to test and compare methods at low cost; (2) challenging simulations that reflect real-world environments and especially can drive perception and planning research; and (3) low-cost platforms with enough software to get started addressing all of these problems. To this end, we propose HomeRobot, a mobile manipulator software stack with associated benchmark in simulation, which is initially based on the low-cost, human-safe Hello Robot Stretch.