Ali Ahmad Malik

Ali Malik, Stephen Mayhew, Chris Piech et al.

We study the problem of controlling the difficulty level of text generated by Large Language Models (LLMs) for contexts where end-users are not fully proficient, such as language learners. Using a novel framework, we evaluate the effectiveness of several key approaches for this task, including few-shot prompting, supervised finetuning, and reinforcement learning (RL), utilising both GPT-4 and open source alternatives like LLama2-7B and Mistral-7B. Our findings reveal a large performance gap between GPT-4 and the open source models when using prompt-based strategies. However, we show how to bridge this gap with a careful combination of finetuning and RL alignment. Our best model, CALM (CEFR-Aligned Language Model), surpasses the performance of GPT-4 and other strategies, at only a fraction of the cost. We further validate the quality of our results through a small-scale human study.

Ali Ahmad Malik, Tariq Masood, Rehana Kousar

As the COVID-19 pandemic expands, the shortening of medical equipment is swelling. A key piece of equipment getting far-out attention has been ventilators. The difference between supply and demand is substantial to be handled with normal production techniques, especially under social distancing measures in place. The study explores the rationale of human-robot teams to ramp up production using advantages of both the ease of integration and maintaining social distancing. The paper presents a model for faster integration of collaborative robots and design guidelines for workstation. The scenarios are evaluated for an open source ventilator through continuous human-robot simulation and amplification of results in a discrete event simulation.

Rishi Bommasani, Drew A. Hudson, Ehsan Adeli et al.

AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.

Ali Ahmad Malik

Objective: The status of human-robot collaboration for assembly applications is reviewed and key current challenges for the research community and practitioners are presented. Background: As the pandemic of COVID-19 started to surface the manufacturers went under pressure to address demand challenges. Social distancing measures made fewer people available to work. In such situations, robots were pointed at to support humans to address a shortage in supply. An important activity where humans are needed in a manufacturing value chain is assembly. HRC assembly systems are supposed to safeguard coexisting humans, perform a range of actions, and often need to be reconfigured to handle product variety. This requires them to be resilient and adaptable to various configurations during their operational life. Besides the potential advantages of using robots the challenges of using them in an industrial assembly are enormous. Methods: This mini-review summarizes the challenges of industrial deployment of collaborative robots for assembly applications. Applications: The documented challenges highlight the future research directions in human-robot interaction for industrial applications.