This paper presents CoLLIE: a simple, yet effective model for continual learning of how language is grounded in vision. Given a pre-trained multimodal embedding model, where language and images are projected in the same semantic space (in this case CLIP by OpenAI), CoLLIE learns a transformation function that adjusts the language embeddings when needed to accommodate new language use. This is done by predicting the difference vector that needs to be applied, as well as a scaling factor for this vector, so that the adjustment is only applied when needed. Unlike traditional few-shot learning, the model does not just learn new classes and labels, but can also generalize to similar language use and leverage semantic compositionality. We verify the model's performance on two different tasks of identifying the targets of referring expressions, where it has to learn new language use. The results show that the model can efficiently learn and generalize from only a few examples, with little interference with the model's original zero-shot performance.

Building autonomous machines that can explore open-ended environments, discover possible interactions and build repertoires of skills is a general objective of artificial intelligence. Developmental approaches argue that this can only be achieved by autotelic agents: intrinsically motivated learning agents that can learn to represent, generate, select and solve their own problems. In recent years, the convergence of developmental approaches with deep reinforcement learning (rl) methods has been leading to the emergence of a new field: developmental reinforcement learning. Developmental rl is concerned with the use of deep rl algorithms to tackle a developmental problem -- the intrinsically motivated acquisition of open-ended repertoires of skills. The self-generation of goals requires the learning of compact goal encodings as well as their associated goal-achievement functions. This raises new challenges compared to standard rl algorithms originally designed to tackle pre-defined sets of goals using external reward signals. The present paper introduces developmental rl and proposes a computational framework based on goal-conditioned rl to tackle the intrinsically motivated skills acquisition problem. It proceeds to present a typology of the various goal representations used in the literature, before reviewing existing methods to learn to represent and prioritize goals in autonomous systems. We finally close the paper by discussing some open challenges in the quest of intrinsically motivated skills acquisition.

Toronto artificial intelligence (AI) startup Ada is bolstering its tech stack with a new machine learning centre in Israel and the appointment of a chief product officer (CPO). "The motivation to open the machine learning centre in Israel stems from the pool of talent there in conversational AI and in machine learning." This week, Ada announced the opening of its office in Israel, where it will be hiring machine learning, engineering, and product teams to continue to develop the conversational AI systems that power its automated brand interaction platform. Israel's growing AI market is what attracted Ada to make inroads into the country, according to the startup. Research firm Tracxn estimates that there are currently 1,100 startups in Israel that use AI as a core component of their offering.

Donna McDaniel of Keller Independent School District in Texas was one of the first teachers in the world to use artificial intelligence-driven coaching for professional learning.

While AI-powered solutions have been in the EdTech industry for some time, the pandemic drastically shifted the landscape, forcing educators/learners to rely on technology more than ever before. AI technologies have the power to optimise both learning and teaching, thus paving the way for the education sector to evolve. Here are a few points on how AI in education can and will shape and define the teaching-learning ecosystem and experience for the future. Routine chores such as grading, evaluating, filing paperwork, making progress reports and organizing resources for lectures will get automated by AI, reducing the time-to-task of teachers and freeing up time for developing students' higher-order thinking and skills. There are already adaptive learning software and digitised programs for students.

The Department of Computing at Imperial College London is a leading department of Computer Science among UK Universities, and has consistently been awarded the highest research rating. In the 2021 Research Excellence Framework (REF) assessment, the Department was ranked 1st out of all UK Computer Science departments and was rated as "Excellent" in the previous national assessment of teaching quality. Imperial was also awarded "Gold" according the last Teaching Excellence Framework (TEF) 2017. The Biomedical Image Analysis (BioMedIA) group is a friendly, vibrant, multi-national team of around 50 academics, post-doctoral researchers and PhD students, pursuing research on novel computational techniques for the processing and analysis of biomedical images. Topics of interest are image registration, reconstruction, segmentation and modelling.

Python is an easy-to-learn programming language, which makes it the most preferred choice for beginners in Data Science, Data Analytics, and Machine Learning. It also has a great community of online learners and excellent data-centric libraries. With so much data being generated, it becomes important that the data we use for Data Science applications like Machine Learning and Predictive Modeling is clean. But what do we mean by clean data? And what makes data dirty in the first place?

Tadej Magajna is a former lead machine learning engineer, former data scientist and now a software engineer at Microsoft. He currently works in a team responsible for language model training and building language packs for keyboards such as Microsoft SwiftKey. He is also a Master of computer science. He started his career as a 15-year-old at a local media company as a web developer and progressed towards more complex engineering and machine learning problems. He tackled problems like NLP market research, public transport bus and train capacity forecasting and finally language model training at his current role.

HCI researchers have been gradually shifting attention from individual users to communities when engaging in research, design, and system development. However, our field has yet to establish a cohesive, systematic understanding of the challenges, benefits, and commitments of community-collaborative approaches to research. We conducted a systematic review and thematic analysis of 47 computing research papers discussing participatory research with communities for the development of technological artifacts and systems, published over the last two decades. From this review, we identified seven themes associated with the evolution of a project: from establishing community partnerships to sustaining results. Our findings suggest that several tensions characterize these projects, many of which relate to the power and position of researchers, and the computing research environment, relative to community partners. We discuss the implications of our findings and offer methodological proposals to guide HCI, and computing research more broadly, towards practices that center communities.

Continual/lifelong learning from a non-stationary input data stream is a cornerstone of intelligence. Despite their phenomenal performance in a wide variety of applications, deep neural networks are prone to forgetting their previously learned information upon learning new ones. This phenomenon is called "catastrophic forgetting" and is deeply rooted in the stability-plasticity dilemma. Overcoming catastrophic forgetting in deep neural networks has become an active field of research in recent years. In particular, gradient projection-based methods have recently shown exceptional performance at overcoming catastrophic forgetting. This paper proposes two biologically-inspired mechanisms based on sparsity and heterogeneous dropout that significantly increase a continual learner's performance over a long sequence of tasks. Our proposed approach builds on the Gradient Projection Memory (GPM) framework. We leverage k-winner activations in each layer of a neural network to enforce layer-wise sparse activations for each task, together with a between-task heterogeneous dropout that encourages the network to use non-overlapping activation patterns between different tasks. In addition, we introduce two new benchmarks for continual learning under distributional shift, namely Continual Swiss Roll and ImageNet SuperDog-40. Lastly, we provide an in-depth analysis of our proposed method and demonstrate a significant performance boost on various benchmark continual learning problems.