Recently, significant progress has been made on Large Vision-Language Models (LVLMs); a new class of VL models that make use of large pre-trained language models. Yet, their vulnerability to Typographic attacks, which involve superimposing misleading text onto an image remain unstudied. Furthermore, prior work typographic attacks rely on sampling a random misleading class from a predefined set of classes. However, the random chosen class might not be the most effective attack. To address these issues, we first introduce a novel benchmark uniquely designed to test LVLMs vulnerability to typographic attacks. Furthermore, we introduce a new and more effective typographic attack: Self-Generated typographic attacks. Indeed, our method, given an image, make use of the strong language capabilities of models like GPT-4V by simply prompting them to recommend a typographic attack. Using our novel benchmark, we uncover that typographic attacks represent a significant threat against LVLM(s). Furthermore, we uncover that typographic attacks recommended by GPT-4V using our new method are not only more effective against GPT-4V itself compared to prior work attacks, but also against a host of less capable yet popular open source models like LLaVA, InstructBLIP, and MiniGPT4.
In this paper, we propose a social robot capable of verbally interacting with children with Autism Spectrum Disorder (ASD). This communication is meant to teach perspective-taking using text generated using a Large Language Model (LLM) pipeline. The social robot NAO acts as a stimulator (verbally describes a social situation and asks a question), prompter (presents three options to choose from), and reinforcer (praises when the answer is correct). For the role of the stimulator, the social situation, questions, and options are generated using our LLM pipeline. We compare two approaches: GPT-2 + BART and GPT-2 + GPT-2, where the first GPT-2 common between the pipelines is used for unsupervised social situation generation. We use the SOCIALIQA dataset to fine-tune all of our LLM pipelines. We found that the GPT-2 + BART pipeline had a better BERTscore for generating the questions and the options by combining their individual loss functions. This observation was also consistent with the human evaluations. Lastly, the unsupervised generation of social situations was visualized using T-SNE plots, and the entire pipeline was evaluated for appropriateness for children with ASD by human experts.
Diffusion models (DMs) are a type of generative model that has a huge impact on image synthesis and beyond. They achieve state-of-the-art generation results in various generative tasks. A great diversity of conditioning inputs, such as text or bounding boxes, are accessible to control the generation. In this work, we propose a conditioning mechanism utilizing Gaussian mixture models (GMMs) as feature conditioning to guide the denoising process. Based on set theory, we provide a comprehensive theoretical analysis that shows that conditional latent distribution based on features and classes is significantly different, so that conditional latent distribution on features produces fewer defect generations than conditioning on classes. Two diffusion models conditioned on the Gaussian mixture model are trained separately for comparison. Experiments support our findings. A novel gradient function called the negative Gaussian mixture gradient (NGMG) is proposed and applied in diffusion model training with an additional classifier. Training stability has improved. We also theoretically prove that NGMG shares the same benefit as the Earth Mover distance (Wasserstein) as a more sensible cost function when learning distributions supported by low-dimensional manifolds.
The development of Courses of Action (COAs) in military operations is traditionally a time-consuming and intricate process. Addressing this challenge, this study introduces COA-GPT, a novel algorithm employing Large Language Models (LLMs) for rapid and efficient generation of valid COAs. COA-GPT incorporates military doctrine and domain expertise to LLMs through in-context learning, allowing commanders to input mission information - in both text and image formats - and receive strategically aligned COAs for review and approval. Uniquely, COA-GPT not only accelerates COA development, producing initial COAs within seconds, but also facilitates real-time refinement based on commander feedback. This work evaluates COA-GPT in a military-relevant scenario within a militarized version of the StarCraft II game, comparing its performance against state-of-the-art reinforcement learning algorithms. Our results demonstrate COA-GPT's superiority in generating strategically sound COAs more swiftly, with added benefits of enhanced adaptability and alignment with commander intentions. COA-GPT's capability to rapidly adapt and update COAs during missions presents a transformative potential for military planning, particularly in addressing planning discrepancies and capitalizing on emergent windows of opportunities.
We consider cooperative semantic text communications facilitated by a relay node. We propose two types of semantic forwarding: semantic lossy forwarding (SLF) and semantic predict-and-forward (SPF). Both are machine learning aided approaches, and, in particular, utilize attention mechanisms at the relay to establish a dynamic semantic state, updated upon receiving a new source signal. In the SLF model, the semantic state is used to decode the received source signal; whereas in the SPF model, it is used to predict the next source signal, enabling proactive forwarding. Our proposed forwarding schemes do not need any channel state information and exhibit consistent performance regardless of the relay's position. Our results demonstrate that the proposed semantic forwarding techniques outperform conventional semantic-agnostic baselines.
Named Entity Recognition (NER) is a sub-task of Natural Language Processing (NLP) that distinguishes entities from unorganized text into predefined categorization. In recent years, a lot of Bangla NLP subtasks have received quite a lot of attention; but Named Entity Recognition in Bangla still lags behind. In this research, we explored the existing state of research in Bangla Named Entity Recognition. We tried to figure out the limitations that current techniques and datasets face, and we would like to address these limitations in our research. Additionally, We developed a Gazetteer that has the ability to significantly boost the performance of NER. We also proposed a new NER solution by taking advantage of state-of-the-art NLP tools that outperform conventional techniques.
Autoregressive Large Language Models (LLMs) trained for next-word prediction have demonstrated remarkable proficiency at producing coherent text. But are they equally adept at forming coherent probability judgments? We use probabilistic identities and repeated judgments to assess the coherence of probability judgments made by LLMs. Our results show that the judgments produced by these models are often incoherent, displaying human-like systematic deviations from the rules of probability theory. Moreover, when prompted to judge the same event, the mean-variance relationship of probability judgments produced by LLMs shows an inverted-U-shaped like that seen in humans. We propose that these deviations from rationality can be explained by linking autoregressive LLMs to implicit Bayesian inference and drawing parallels with the Bayesian Sampler model of human probability judgments.
The recent burst in Large Language Models has opened new frontiers in human-like text processing and generation. However, alongside their remarkable growth, Large Language Models have encountered critical challenges including issues of hallucination, bias, real-time knowledge updates, and the high costs of implementation and maintenance in commercial settings. Vector Databases, another increasingly popular tool, offer potential solutions to these challenges. These databases are adept at handling high-dimensional data and are crucial for tasks such as efficient information retrieval and semantic search. By integrating with Large Language Models, they significantly enhance AI systems' ability to manage and utilize diverse data more effectively. This survey paper provides an in-depth and unique analysis of the intersection between Large Language Models and Vector Databases.
In this paper, we propose a novel cascaded diffusion-based generative framework for text-driven human motion synthesis, which exploits a strategy named GradUally Enriching SyntheSis (GUESS as its abbreviation). The strategy sets up generation objectives by grouping body joints of detailed skeletons in close semantic proximity together and then replacing each of such joint group with a single body-part node. Such an operation recursively abstracts a human pose to coarser and coarser skeletons at multiple granularity levels. With gradually increasing the abstraction level, human motion becomes more and more concise and stable, significantly benefiting the cross-modal motion synthesis task. The whole text-driven human motion synthesis problem is then divided into multiple abstraction levels and solved with a multi-stage generation framework with a cascaded latent diffusion model: an initial generator first generates the coarsest human motion guess from a given text description; then, a series of successive generators gradually enrich the motion details based on the textual description and the previous synthesized results. Notably, we further integrate GUESS with the proposed dynamic multi-condition fusion mechanism to dynamically balance the cooperative effects of the given textual condition and synthesized coarse motion prompt in different generation stages. Extensive experiments on large-scale datasets verify that GUESS outperforms existing state-of-the-art methods by large margins in terms of accuracy, realisticness, and diversity. Code is available at https://github.com/Xuehao-Gao/GUESS.
In this paper, we introduce a novel and simple method for obtaining high-quality text embeddings using only synthetic data and less than 1k training steps. Unlike existing methods that often depend on multi-stage intermediate pre-training with billions of weakly-supervised text pairs, followed by fine-tuning with a few labeled datasets, our method does not require building complex training pipelines or relying on manually collected datasets that are often constrained by task diversity and language coverage. We leverage proprietary LLMs to generate diverse synthetic data for hundreds of thousands of text embedding tasks across nearly 100 languages. We then fine-tune open-source decoder-only LLMs on the synthetic data using standard contrastive loss. Experiments demonstrate that our method achieves strong performance on highly competitive text embedding benchmarks without using any labeled data. Furthermore, when fine-tuned with a mixture of synthetic and labeled data, our model sets new state-of-the-art results on the BEIR and MTEB benchmarks.