We present SOLO, a single transformer for Scalable visiOn-Language mOdeling. Current large vision-language models (LVLMs) such as LLaVA mostly employ heterogeneous architectures that connect pre-trained visual encoders with large language models (LLMs) to facilitate visual recognition and complex reasoning. Although achieving remarkable performance with relatively lightweight training, we identify four primary scalability limitations: (1) The visual capacity is constrained by pre-trained visual encoders, which are typically an order of magnitude smaller than LLMs. (2) The heterogeneous architecture complicates the use of established hardware and software infrastructure. (3) Study of scaling laws on such architecture must consider three separate components - visual encoder, connector, and LLMs, which complicates the analysis. (4) The use of existing visual encoders typically requires following a pre-defined specification of image inputs pre-processing, for example, by reshaping inputs to fixed-resolution square images, which presents difficulties in processing and training on high-resolution images or those with unusual aspect ratio. A unified single Transformer architecture, like SOLO, effectively addresses these scalability concerns in LVLMs; however, its limited adoption in the modern context likely stems from the absence of reliable training recipes that balance both modalities and ensure stable training for billion-scale models. In this paper, we introduce the first open-source training recipe for developing SOLO, an open-source 7B LVLM using moderate academic resources. The training recipe involves initializing from LLMs, sequential pre-training on ImageNet and web-scale data, and instruction fine-tuning on our curated high-quality datasets. On extensive evaluation, SOLO demonstrates performance comparable to LLaVA-v1.5-7B, particularly excelling in visual mathematical reasoning.

While large language models (LLMs) have demonstrated superior multi-task capabilities, understanding the learning mechanisms behind this is still a challenging problem. In this paper, we attempt to understand such mechanisms from the perspective of neurons. Specifically, we detect task-sensitive neurons in LLMs via gradient attribution on task-specific data. Through extensive deactivation and fine-tuning experiments, we demonstrate that the detected neurons are highly correlated with the given task, which we term as task-specific neurons. With these identified task-specific neurons, we delve into two common problems in multi-task learning and continuous learning: Generalization and Catastrophic Forgetting. We find that the overlap of task-specific neurons is strongly associated with generalization and specialization across tasks. Interestingly, at certain layers of LLMs, there is a high similarity in the parameters of different task-specific neurons, and such similarity is highly correlated with the generalization performance. Inspired by these findings, we propose a neuron-level continuous fine-tuning method that only fine-tunes the current task-specific neurons during continuous learning, and extensive experiments demonstrate the effectiveness of the proposed method. Our study provides insights into the interpretability of LLMs in multi-task learning.

Satellite imagery has emerged as an important tool to analyse demographic, health, and development indicators. While various deep learning models have been built for these tasks, each is specific to a particular problem, with few standard benchmarks available. We propose a new dataset pairing satellite imagery and high-quality survey data on child poverty to benchmark satellite feature representations. Our dataset consists of 33,608 images, each 10 km $\times$ 10 km, from 19 countries in Eastern and Southern Africa in the time period 1997-2022. As defined by UNICEF, multidimensional child poverty covers six dimensions and it can be calculated from the face-to-face Demographic and Health Surveys (DHS) Program . As part of the benchmark, we test spatial as well as temporal generalization, by testing on unseen locations, and on data after the training years. Using our dataset we benchmark multiple models, from low-level satellite imagery models such as MOSAIKS , to deep learning foundation models, which include both generic vision models such as Self-Distillation with no Labels (DINOv2) models and specific satellite imagery models such as SatMAE. We provide open source code for building the satellite dataset, obtaining ground truth data from DHS and running various models assessed in our work.

Although attention mechanisms have achieved considerable progress in Transformer-based architectures across various Artificial Intelligence (AI) domains, their inner workings remain to be explored. Existing explainable methods have different emphases but are rather one-sided. They primarily analyse the attention mechanisms or gradient-based attribution while neglecting the magnitudes of input feature values or the skip-connection module. Moreover, they inevitably bring spurious noisy pixel attributions unrelated to the model's decision, hindering humans' trust in the spotted visualization result. Hence, we propose an easy-to-implement but effective way to remedy this flaw: Smooth Noise Norm Attention (SNNA). We weigh the attention by the norm of the transformed value vector and guide the label-specific signal with the attention gradient, then randomly sample the input perturbations and average the corresponding gradients to produce noise-free attribution. Instead of evaluating the explanation method on the binary or multi-class classification tasks like in previous works, we explore the more complex multi-label classification scenario in this work, i.e., the driving action prediction task, and trained a model for it specifically. Both qualitative and quantitative evaluation results show the superiority of SNNA compared to other SOTA attention-based explainable methods in generating a clearer visual explanation map and ranking the input pixel importance.

The Music Emotion Recognition (MER) field has seen steady developments in recent years, with contributions from feature engineering, machine learning, and deep learning. The landscape has also shifted from audio-centric systems to bimodal ensembles that combine audio and lyrics. However, a severe lack of public and sizeable bimodal databases has hampered the development and improvement of bimodal audio-lyrics systems. This article proposes three new audio, lyrics, and bimodal MER research datasets, collectively called MERGE, created using a semi-automatic approach. To comprehensively assess the proposed datasets and establish a baseline for benchmarking, we conducted several experiments for each modality, using feature engineering, machine learning, and deep learning methodologies. In addition, we propose and validate fixed train-validate-test splits. The obtained results confirm the viability of the proposed datasets, achieving the best overall result of 79.21% F1-score for bimodal classification using a deep neural network.

The integration of artificial intelligence (AI) chatbots into higher education marks a shift towards a new generation of pedagogical tools, mirroring the arrival of milestones like the internet. With the launch of ChatGPT-4 Turbo in November 2023, we developed a ChatGPT-based teaching application (https://chat.openai.com/g/g-1imx1py4K-chatge-medical-imaging) and integrated it into our undergraduate medical imaging course in the Spring 2024 semester. This study investigates the use of ChatGPT throughout a semester-long trial, providing insights into students' engagement, perception, and the overall educational effectiveness of the technology. We systematically collected and analyzed data concerning students' interaction with ChatGPT, focusing on their attitudes, concerns, and usage patterns. The findings indicate that ChatGPT offers significant advantages such as improved information access and increased interactivity, but its adoption is accompanied by concerns about the accuracy of the information provided and the necessity for well-defined guidelines to optimize its use.

While imitation learning provides a simple and effective framework for policy learning, acquiring consistent action during robot execution remains a challenging task. Existing approaches primarily focus on either modifying the action representation at data curation stage or altering the model itself, both of which do not fully address the scalability of consistent action generation. To overcome this limitation, we introduce the Diff-Control policy, which utilizes a diffusion-based model to learn action representation from a state-space modeling viewpoint. We demonstrate that diffusion-based policies can acquire statefulness through a Bayesian formulation facilitated by ControlNet, leading to improved robustness and success rates. Our experimental results demonstrate the significance of incorporating action statefulness in policy learning, where Diff-Control shows improved performance across various tasks. Specifically, Diff-Control achieves an average success rate of 72% and 84% on stateful and dynamic tasks, respectively. Notably, Diff-Control also shows consistent performance in the presence of perturbations, outperforming other state-of-the-art methods that falter under similar conditions. Project page: https://diff-control.github.io/

This paper will examine the cognitive processes involved in palpation in order to develop an appropriate remote palpation system. In a conventional remote palpation system, the tactile condition of the patient is conveyed to the doctors using a force feedback system. A clarification of the cognitive process during palpation suggests that the purpose of palpation is to formulate a clear idea about the patient's medical problems using the tactile sensation as a trigger to combine the results of other assessments, past experience and memory, and patient reactions to the doctor's touch. This is in contrast to the objective of acquiring the detailed tactile condition of the affected body part. In order to demonstrate this purpose, we will describe the two significant signal pathways for the perception of tactile sensation, both in doctors and patients. The perception of doctors progresses as the result of active touch to the affected part, thereby implying that the simultaneous stimulation of kinaesthetic and tactile sensation is necessary. Conversely, the tactile sensation experienced by patients is the result of passive touch, which evokes a more subjective and emotional response. Patients both explicitly and implicitly perceive the stimulation, and doctors use these perceptions as reactions of the pain to the doctors' touch. This paper proposes the fundamental concept of a remote palpation system, ``Palpation Reality beyond Real'', to achieve the purpose of palpation. Palpation reality implies a system in which the whole cognitive process progresses at the same level or better than palpation in the standard examination, rather than presenting the real tactile sensation.

The Large language models (LLMs) have showcased superior capabilities in sophisticated tasks across various domains, stemming from basic question-answer (QA), they are nowadays used as decision assistants or explainers for unfamiliar content. However, they are not always correct due to the data sparsity in specific domain corpus, or the model's hallucination problems. Given this, how much should we trust the responses from LLMs? This paper presents a novel way to evaluate the uncertainty that captures the directional instability, by constructing a directional graph from entailment probabilities, and we innovatively conduct Random Walk Laplacian given the asymmetric property of a constructed directed graph, then the uncertainty is aggregated by the derived eigenvalues from the Laplacian process. We also provide a way to incorporate the existing work's semantics uncertainty with our proposed layer. Besides, this paper identifies the vagueness issues in the raw response set and proposes an augmentation approach to mitigate such a problem, we conducted extensive empirical experiments and demonstrated the superiority of our proposed solutions.

Large Language Models (LLMs) have shown promise in Automated Essay Scoring (AES), but their zero-shot and few-shot performance often falls short compared to state-of-the-art models and human raters. However, fine-tuning LLMs for each specific task is impractical due to the variety of essay prompts and rubrics used in real-world educational contexts. This study proposes a novel approach combining LLMs and Comparative Judgment (CJ) for AES, using zero-shot prompting to choose between two essays. We demonstrate that a CJ method surpasses traditional rubric-based scoring in essay scoring using LLMs.