This paper introduces an unobtrusive in-situ measurement method to detect user behavior changes during arbitrary exposures in XR systems. Here, such behavior changes are typically associated with the Proteus effect or bodily affordances elicited by different avatars that the users embody in XR. We present a biometric user model based on deep metric similarity learning, which uses high-dimensional embeddings as reference vectors to identify behavior changes of individual users. We evaluate our model against two alternative approaches: a (non-learned) motion analysis based on central tendencies of movement patterns and subjective post-exposure embodiment questionnaires frequently used in various XR exposures. In a within-subject study, participants performed a fruit collection task while embodying avatars of different body heights (short, actual-height, and tall). Subjective assessments confirmed the effective manipulation of perceived body schema, while the (non-learned) objective analyses of head and hand movements revealed significant differences across conditions. Our similarity learning model trained on the motion data successfully identified the elicited behavior change for various query and reference data pairings of the avatar conditions. The approach has several advantages in comparison to existing methods: 1) In-situ measurement without additional user input, 2) generalizable and scalable motion analysis for various use cases, 3) user-specific analysis on the individual level, and 4) with a trained model, users can be added and evaluated in real time to study how avatar changes affect behavior.

--Membership inference attacks (MIAs) aim to determine whether specific data were used to train a model. While extensively studied on classification models, their impact on time series forecasting remains largely unexplored. We address this gap by introducing two new attacks: (i) an adaptation of multivariate LiRA, a state-of-the-art MIA originally developed for classification models, to the time-series forecasting setting, and (ii) a novel end-to-end learning approach called Deep Time Series (DTS) attack. We benchmark these methods against adapted versions of other leading attacks from the classification setting. We evaluate all attacks in realistic settings on the TUH-EEG and ELD datasets, targeting two strong forecasting architectures, LSTM and the state-of-the-art N-HiTS, under both record-and user-level threat models. Our results show that forecasting models are vulnerable, with user-level attacks often achieving perfect detection. The proposed methods achieve the strongest performance in several settings, establishing new baselines for privacy risk assessment in time series forecasting. Furthermore, vulnerability increases with longer prediction horizons and smaller training populations, echoing trends observed in large language models. The integration of machine learning (ML) into real-world applications has significantly expanded the ways in which industries use data for decision-making. Time-series forecasting, in particular, is widely applied in domains such as healthcare, finance, and smart infrastructure. However, when deployed in settings that involve sensitive information, these models raise important concerns regarding data privacy. Modern ML models have been shown to memorize information from their training data [1].

The generation of data is a common approach to improve the performance of machine learning tasks, among which is the training of models for classification. In this paper, we present TAGAL, a collection of methods able to generate synthetic tabular data using an agentic workflow. The methods leverage Large Language Models (LLMs) for an automatic and iterative process that uses feedback to improve the generated data without any further LLM training. The use of LLMs also allows for the addition of external knowledge in the generation process. We evaluate TAGAL across diverse datasets and different aspects of quality for the generated data. We look at the utility of downstream ML models, both by training classifiers on synthetic data only and by combining real and synthetic data. Moreover, we compare the similarities between the real and the generated data. We show that TAGAL is able to perform on par with state-of-the-art approaches that require LLM training and generally outperforms other training-free approaches.

In this paper, we introduce Technical-Embeddings, a novel framework designed to optimize semantic retrieval in technical documentation, with applications in both hardware and software development. Our approach addresses the challenges of understanding and retrieving complex technical content by leveraging the capabilities of Large Language Models (LLMs). First, we enhance user queries by generating expanded representations that better capture user intent and improve dataset diversity, thereby enriching the fine-tuning process for embedding models. Second, we apply summary extraction techniques to encode essential contextual information, refining the representation of technical documents. To further enhance retrieval performance, we fine-tune a bi-encoder BERT model using soft prompting, incorporating separate learning parameters for queries and document context to capture fine-grained semantic nuances. We evaluate our approach on two public datasets, RAG-EDA and Rust-Docs-QA, demonstrating that Technical-Embeddings significantly outperforms baseline models in both precision and recall. Our findings highlight the effectiveness of integrating query expansion and contextual summarization to enhance information access and comprehension in technical domains. This work advances the state of Retrieval-Augmented Generation (RAG) systems, offering new avenues for efficient and accurate technical document retrieval in engineering and product development workflows.

Recent advances in AI raise the possibility that AI systems will one day be able to do anything humans can do, only better. If artificial general intelligence (AGI) is achieved, AI systems may be able to understand, reason, problem solve, create, and evolve at a level and speed that humans will increasingly be unable to match, or even understand. These possibilities raise a natural question as to whether AI will eventually become superior to humans, a successor "digital species", with a rightful claim to assume leadership of the universe. However, a deeper consideration suggests the overlooked differentiator between human beings and AI is not the brain, but the central nervous system (CNS), providing us with an immersive integration with physical reality. It is our CNS that enables us to experience emotion including pain, joy, suffering, and love, and therefore to fully appreciate the consequences of our actions on the world around us. And that emotional understanding of the consequences of our actions is what is required to be able to develop sustainable ethical systems, and so be fully qualified to be the leaders of the universe. A CNS cannot be manufactured or simulated; it must be grown as a biological construct. And so, even the development of consciousness will not be sufficient to make AI systems superior to humans. AI systems may become more capable than humans on almost every measure and transform our society. However, the best foundation for leadership of our universe will always be DNA, not silicon.

Systemic Functional Grammar and its branch, Cardiff Grammar, have been widely applied to discourse analysis, semantic function research, and other tasks across various languages and texts. However, an automatic annotation system based on this theory for Chinese texts has not yet been developed, which significantly constrains the application and promotion of relevant theories. To fill this gap, this research introduces a functional syntax annotation model for Chinese based on RoBERTa (Robustly Optimized BERT Pretraining Approach). The study randomly selected 4,100 sentences from the People's Daily 2014 corpus and annotated them according to functional syntax theory to establish a dataset for training. The study then fine-tuned the RoBERTa-Chinese wwm-ext model based on the dataset to implement the named entity recognition task, achieving an F1 score of 0.852 on the test set that significantly outperforms other comparative models. The model demonstrated excellent performance in identifying core syntactic elements such as Subject (S), Main Verb (M), and Complement (C). Nevertheless, there remains room for improvement in recognizing entities with imbalanced label samples. As the first integration of functional syntax with attention-based NLP models, this research provides a new method for automated Chinese functional syntax analysis and lays a solid foundation for subsequent studies.

Due to their powerful feature association capabilities, neural network-based computer vision models have the ability to detect and exploit unintended patterns within the data, potentially leading to correct predictions based on incorrect or unintended but statistically relevant signals. These clues may vary from simple color aberrations to small texts within the image. In situations where these unintended signals align with the predictive task, models can mistakenly link these features with the task and rely on them for making predictions. This phenomenon is referred to as spurious correlations, where patterns appear to be associated with the task but are actually coincidental. As a result, detection and mitigation of spurious correlations have become crucial tasks for building trustworthy, reliable, and generalizable machine learning models. In this work, we present a novel method to detect spurious correlations in vision transformers, a type of neural network architecture that gained significant popularity in recent years. Using both supervised and self-supervised trained models, we present large-scale experiments on the ImageNet dataset demonstrating the ability of the proposed method to identify spurious correlations. We also find that, even if the same architecture is used, the training methodology has a significant impact on the model's reliance on spurious correlations. Furthermore, we show that certain classes in the ImageNet dataset contain spurious signals that are easily detected by the models and discuss the underlying reasons for those spurious signals. In light of our findings, we provide an exhaustive list of the aforementioned images and call for caution in their use in future research efforts. Lastly, we present a case study investigating spurious signals in invasive breast mass classification, grounding our work in real-world scenarios.

-- We propose Spatial-Aware Correlated Multiple Instance Learning (SAC-MIL) for performing WSI classification. SAC-MIL consists of a positional encoding module to encode position information and a SAC block to perform full instance correlations. The positional encoding module utilizes the instance coordinates within the slide to encode the spatial relationships instead of the instance index in the input WSI sequence. The positional encoding module can also handle the length extrapolation issue where the training and testing sequences have different lengths. The SAC block is an MLP-based method that performs full instance correlation in linear time complexity with respect to the sequence length. Due to the simple structure of MLP, it is easy to deploy since it does not require custom CUDA kernels, compared to Transformer-based methods for WSI classification. SAC-MIL has achieved state-of-the-art performance on the CAMEL YON-16, TCGA-LUNG, and TCGA-BRAC datasets. The code will be released upon acceptance. Index Terms -- Enter about five key words or phrases in alphabetical order, separated by commas. Whole Slide Image (WSI) classification plays a critical role in computer-aided diagnosis (CAD) [12]. WSI classification aims to perform tumor classification of a given slide.

We introduce Llama-3-Motif, a language model consisting of 102 billion parameters, specifically designed to enhance Korean capabilities while retaining strong performance in English. Developed on the Llama 3 architecture, Llama-3-Motif employs advanced training techniques, including LlamaPro and Masked Structure Growth, to effectively scale the model without altering its core Transformer architecture. Using the MoAI platform for efficient training across hyperscale GPU clusters, we optimized Llama-3-Motif using a carefully curated dataset that maintains a balanced ratio of Korean and English data. Llama-3-Motif shows decent performance on Korean-specific benchmarks, outperforming existing models and achieving results comparable to GPT-4.

The effectiveness of Large Language Models (LLMs) diminishes for extremely low-resource languages, such as indigenous languages, primarily due to the lack of labeled data. Despite growing interest, the availability of high-quality natural language processing (NLP) datasets for these languages remains limited, making it difficult to develop robust language technologies. This paper addresses such gap by focusing on Ladin, an endangered Romance language, specifically targeting the Val Badia variant. Leveraging a small set of parallel Ladin-Italian sentence pairs, we create synthetic datasets for sentiment analysis and multiple-choice question answering (MCQA) by translating monolingual Italian data. To ensure linguistic quality and reliability, we apply rigorous filtering and back-translation procedures in our method. We further demonstrate that incorporating these synthetic datasets into machine translation training leads to substantial improvements over existing Italian-Ladin translation baselines. Our contributions include the first publicly available sentiment analysis and MCQA datasets for Ladin, establishing foundational resources that can support broader NLP research and downstream applications for this underrepresented language.