Large Language Models (LLMs) have demonstrated strong capabilities in transforming text descriptions or tables to data visualizations via instruction-tuning methods. However, it is not straightforward to apply these methods directly for a more real-world use case of visualizing data from long documents based on user-given intents, as opposed to the user pre-selecting the relevant content manually. We introduce the task of intent-based chart generation from documents: given a user-specified intent and document(s), the goal is to generate a chart adhering to the intent and grounded on the document(s) in a zero-shot setting. We propose an unsupervised, two-staged framework in which an LLM first extracts relevant information from the document(s) by decomposing the intent and iteratively validates and refines this data. Next, a heuristic-guided module selects an appropriate chart type before final code generation. To assess the data accuracy of the generated charts, we propose an attribution-based metric that uses a structured textual representation of charts, instead of relying on visual decoding metrics that often fail to capture the chart data effectively. To validate our approach, we curate a dataset comprising of 1,242 $<$intent, document, charts$>$ tuples from two domains, finance and scientific, in contrast to the existing datasets that are largely limited to parallel text descriptions/ tables and their corresponding charts. We compare our approach with baselines using single-shot chart generation using LLMs and query-based retrieval methods; our method outperforms by upto $9$ points and $17$ points in terms of chart data accuracy and chart type respectively over the best baselines.

Machine unlearning aims to remove knowledge of the specific training data in a well-trained model. Currently, machine unlearning methods typically handle all forgetting data in a single batch, removing the corresponding knowledge all at once upon request. However, in practical scenarios, requests for data removal often arise in a streaming manner rather than in a single batch, leading to reduced efficiency and effectiveness in existing methods. Such challenges of streaming forgetting have not been the focus of much research. In this paper, to address the challenges of performance maintenance, efficiency, and data access brought about by streaming unlearning requests, we introduce a streaming unlearning paradigm, formalizing the unlearning as a distribution shift problem. We then estimate the altered distribution and propose a novel streaming unlearning algorithm to achieve efficient streaming forgetting without requiring access to the original training data. Theoretical analyses confirm an $O(\sqrt{T} + V_T)$ error bound on the streaming unlearning regret, where $V_T$ represents the cumulative total variation in the optimal solution over $T$ learning rounds. This theoretical guarantee is achieved under mild conditions without the strong restriction of convex loss function. Experiments across various models and datasets validate the performance of our proposed method.

Federated learning is a promising privacy-preserving paradigm for distributed machine learning. In this context, there is sometimes a need for a specialized process called machine unlearning, which is required when the effect of some specific training samples needs to be removed from a learning model due to privacy, security, usability, and/or legislative factors. However, problems arise when current centralized unlearning methods are applied to existing federated learning, in which the server aims to remove all information about a class from the global model. Centralized unlearning usually focuses on simple models or is premised on the ability to access all training data at a central node. However, training data cannot be accessed on the server under the federated learning paradigm, conflicting with the requirements of the centralized unlearning process. Additionally, there are high computation and communication costs associated with accessing clients' data, especially in scenarios involving numerous clients or complex global models. To address these concerns, we propose a more effective and efficient federated unlearning scheme based on the concept of model explanation. Model explanation involves understanding deep networks and individual channel importance, so that this understanding can be used to determine which model channels are critical for classes that need to be unlearned. We select the most influential channels within an already-trained model for the data that need to be unlearned and fine-tune only influential channels to remove the contribution made by those data. In this way, we can simultaneously avoid huge consumption costs and ensure that the unlearned model maintains good performance. Experiments with different training models on various datasets demonstrate the effectiveness of the proposed approach.

Introduction: The use of a wide range of computer vision solutions, and more recently high-end Inertial Measurement Units (IMU) have become increasingly popular for assessing human physical activity in clinical and research settings [1]. Nevertheless, to increase the feasibility of patient tracking in out-of-the-lab settings, it is necessary to use a reduced number of devices for movement acquisition. Promising solutions in this context are IMU-based wearables and single camera systems [2]. Additionally, the development of machine learning systems able to recognize and digest clinically relevant data in-the-wild is needed, and therefore determining the ideal input to those is crucial [3]. Research question: For upper-limb activity recognition out-of-the-lab, do wearables or single camera offer better performance?

Abstract. While the use of location-based augmented reality (AR) for education has demonstrated benefits on participants' motivation, engagement, and on their physical activity, geolocation data inaccuracy causes augmented objects to jitter or drift, which is a factor in downgrading user experience. We developed a free and open source web AR application and conducted a comparative user test (n = 54) in order to assess the impact of geolocation data on usability, exploration, and focus. A control group explored biodiversity in nature using the system in combination with embedded GNSS data, and an experimental group used an external module for RTK data. During the test, eye tracking data, geolocated traces, and in-app user-triggered events were recorded. Participants answered usability questionnaires (SUS, UEQ, HARUS).We found that the geolocation data the RTK group was exposed to was less accurate in average than that of the control group. The RTK group reported lower usability scores on all scales, of which 5 out of 9 were significant, indicating that inaccurate data negatively predicts usability. The GNSS group walked more than the RTK group, indicating a partial effect on exploration. We found no significant effect on interaction time with the screen, indicating no specific relation between data accuracy and focus. While RTK data did not allow us to better the usability of location-based AR interfaces, results allow us to assess our system's overall usability as excellent, and to define optimal operating conditions for future use with pupils.

The Italian privacy watchdog mandated a ban on the popular chatbot ChatGPT and launched an investigation on its provider OpenAI for suspected breaches of EU data protection rules. Italy's Garante for the protection of personal data on Friday (31 March) accused the AI system of breaching the EU General Data Protection Regulation (GDPR) and failing to implement age verification systems. The blocking of the site for Italian users is temporary and will last until the provider OpenAI respects the EU privacy framework when processing the personal data of Italian users. The Italian data protection authority has also initiated an investigation into the American tech company. Launched in November, ChatGPT has been notorious for its unprecedented ability to generate human-like text based on prompts.

We propose a circuit-level backdoor attack, \textit{QTrojan}, against Quantum Neural Networks (QNNs) in this paper. QTrojan is implemented by few quantum gates inserted into the variational quantum circuit of the victim QNN. QTrojan is much stealthier than a prior Data-Poisoning-based Backdoor Attack (DPBA), since it does not embed any trigger in the inputs of the victim QNN or require the access to original training datasets. Compared to a DPBA, QTrojan improves the clean data accuracy by 21\% and the attack success rate by 19.9\%.

The push to deliver unbiased and responsible artificial intelligence is admirable, but there are many roadblocks to overcome. Chiefly, AI is only as fair as the data that goes into it. In light of the slow progress addressing AI bias and unfairness, business and technology leaders may be finally arriving at a consensus that they need to concentrate on more "responsible" approaches to AI. A recent survey of 504 IT executives, released by Appen and conducted by The Harris Poll, finds heightened concern about the data that is increasingly driving decisions about customers, markets, and opportunities. It also hints at recognition by both types of leaders that the data they have tends to be problematic, wreaking damage to people, communities, and businesses.

Data is the lifeblood of machine. But organizations continue to struggle to obtain good, clean data to sustain their AI and machine learning initiatives, according to Appen's State of AI and Machine Learning report published this week. Of the four stages of AI–data sourcing, data preparation, model training and deployment, and human-guided model evaluation–data sourcing consumes the most resources, takes the most time, and is the most challenging, according to Appen's survey of 504 business leader and technologists. On average, data sourcing consumes 34% of an organization's AI budget, versus 24% each for data preparation and model testing and deployment and 15% for model evaluation, according to Appen's survey, which was conducted by the Harris Poll and included IT decision makers, business leaders and managers, and technical practitioners from the US, UK, Ireland, and Germany. Finally, 42% of technologists find data sourcing to be the most challenging stage of AI lifecycle, compared to model evaluation (41%), model testing and deployment (38%) and data preparation (34%).

There is so much buzz about artificial intelligence (AI) and machine learning today. It is no longer surprising to realize that most of the tools you use online, from your smartphones, most websites, and various devices, use AI-powered machine learning to enhance your interaction with multiple applications. Some machine learning applications include facial recognition, speech recognition, financial security, bus schedules, traffic prediction, medical services, social media, customer support, and retail. Moreover, writing tools such as Spell Check are developed using machine learning. Another excellent use of machine learning applications is predictive analytics.