The generalization of language models (LMs) is undergoing active debates, contrasting their potential for general intelligence with their struggles with basic knowledge composition (e.g., reverse/transition curse). This paper uncovers the phenomenon of linear correlations in LMs during knowledge composition. For explanation, there exists a linear transformation between certain related knowledge that maps the next token prediction logits from one prompt to another, e.g., "X lives in the city of" $\rightarrow$ "X lives in the country of" for every given X. This mirrors the linearity in human knowledge composition, such as Paris $\rightarrow$ France. Our findings indicate that the linear transformation is resilient to large-scale fine-tuning, generalizing updated knowledge when aligned with real-world relationships, but causing hallucinations when it deviates. Empirical results suggest that linear correlation can serve as a potential identifier of LM's generalization. Finally, we show such linear correlations can be learned with a single feedforward network and pre-trained vocabulary representations, indicating LM generalization heavily relies on the latter.

State of the art Named Entity Recognition (NER) models have achieved an impressive ability to extract common phrases from text that belong to labels such as location, organization, time, and person. However, typical NER systems that rely on having seen a specific entity in their training data in order to label an entity perform poorly on rare or unseen entities ta in order to label an entity perform poorly on rare or unseen entities (Derczynski et al., 2017). This paper attempts to improve recognition of person names, a diverse category that can grow any time someone is born or changes their name. In order for downstream tasks to not exhibit bias based on cultural background, a model should perform well on names from a variety of backgrounds. In this paper I experiment with the training data and input structure of an English Bi-LSTM name recognition model. I look at names from 103 countries to compare how well the model performs on names from different cultures, specifically in the context of a downstream task where extracted names will be matched to information on file. I find that a model with combined character and word input outperforms word-only models and may improve on accuracy compared to classical NER models that are not geared toward identifying unseen entity values.

The 11th International Conference on Learning Representations (ICLR) is taking place this week in Kigali, Rwanda, the first time a major AI conference has taken place in-person in Africa. The program includes workshops, contributed talks, affinity group events, and socials. In addition, a total of six invited talks covered a broad range of topics. In this post we give a flavour of the first two of these presentations. Sofia Crepso is an artist who explores the interaction between biological systems and AI.

The team does so by providing real time models and batch applications in the realm of risk and financial crime over the whole SumUp life cycle and products. Together with the risk platform squad the risk modelling team builds the necessary platform foundations for scalable and reliable ML model serving and development in SumUp. The platform enables a global approach supported by local specifics. Are you up for the challenge? At SumUp, we are driven to empower small businesses across the globe by de-hassling their lives and helping them to succeed.

Experian is the world's leading global information services company. During life's big moments -- from buying a home or a car to sending a child to college to growing a business by connecting with new customers -- we empower consumers and our clients to manage their data with confidence. We have 20,000 people operating across 44 countries. By investing in our people, technology and innovation, we can help transform businesses, help communities prosper, enable more people to feel included in the financial opportunities that should be available to them, and help people to thrive. We're looking for inspired employees that want to make an impact on people and business.

The 21 cm spectral line emission of atomic neutral hydrogen (HI) is one of the primary wavelengths observed in radio astronomy. However, the signal is intrinsically faint and the HI content of galaxies depends on the cosmic environment, requiring large survey volumes and survey depth to investigate the HI Universe. As the amount of data coming from these surveys continues to increase with technological improvements, so does the need for automatic techniques for identifying and characterising HI sources while considering the tradeoff between completeness and purity. This study aimed to find the optimal pipeline for finding and masking the most sources with the best mask quality and the fewest artefacts in 3D neutral hydrogen cubes. Various existing methods were explored in an attempt to create a pipeline to optimally identify and mask the sources in 3D neutral hydrogen 21 cm spectral line data cubes. Two traditional source-finding methods were tested, SoFiA and MTObjects, as well as a new supervised deep learning approach, in which a 3D convolutional neural network architecture, known as V-Net was used. These three source-finding methods were further improved by adding a classical machine learning classifier as a post-processing step to remove false positive detections. The pipelines were tested on HI data cubes from the Westerbork Synthesis Radio Telescope with additional inserted mock galaxies. SoFiA combined with a random forest classifier provided the best results, with the V-Net-random forest combination a close second. We suspect this is due to the fact that there are many more mock sources in the training set than real sources. There is, therefore, room to improve the quality of the V-Net network with better-labelled data such that it can potentially outperform SoFiA.

Abstract-- This paper presents a serial chain hinge support, a rigid but flexible structure that improves the mechanical performance and robustness of soft-fingered grippers. Gravity can reduce the integrity of soft fingers in horizontal approach, resulting in lower maximum payload caused by a large deflection of fingers. To substantiate our claim we performed several experiments on payload and deflection of the SofIA gripper under both horizontal and vertical approach. In addition, we show that this reinforcement does not impede the original compliant behavior of the gripper, maintaining the original kinematic model functionality. Finally, we validated the improved SofIA gripper in agricultural and everyday activities.

First and foremost, let us understand the meaning of "telehealth." The word'tele' means "distance" and'health' means "to heal". Telemedicine also refers to the practice of medicine at a distance whereby information technology is used to ensure the delivery of medical care services. By using mobile phones, laptops, and computers, healthcare providers and doctors can communicate with their patients virtually and write prescriptions or follow-ups. But, at the same time, with the rise of innovative technologies and the use of AI in healthcare -- healthcare businesses have taken a different shape, from traditional styles to telehealth.

Data Scientists and analysts work to constantly deliver valuable insights from data. In many cases, these individuals practice a Code First approach, using a programming language like R or Python to explore and understand data. Once an analysis reaches conclusion, it is important to carefully consider what happens next. Perhaps the analysis resulted in a complex machine learning model that can generate valuable predictions on new data. Or perhaps it resulted in some new business logic that can be implemented to improve efficiency.

The moon's shadowed, frigid nooks and crannies may hold frozen water in more places and in larger quantities than previously suspected. And for the first time, the presence of water on the moon's sunlit surface has been confirmed, scientists reported Monday. That's good news for astronauts at future lunar bases who could tap into these resources for drinking and making rocket fuel. While previous observations have indicated millions of tons of ice in the permanently shadowed craters of the moon's poles, a pair of studies in the journal Nature Astronomy take the availability of lunar surface water to a new level. More than 15,400 square miles of lunar terrain have the capability to trap water in the form of ice, according to a team led by the University of Colorado's Paul Hayne.