Visually rich documents (VRDs) such as invoices, utility bills, and insurance quotes present unique challenges in terms of information extraction (IE). The varied layouts and formats, coupled with both textual and visual properties, require complex, resource-intensive solutions. Many existing strategies rely on supervised learning, which necessitates a vast pool of human-labeled training samples. This not…
Researchers from the Shanghai AI Laboratory and TapTap have developed a Linear Attention Sequence Parallel (LASP) technique that optimizes sequence parallelism on linear transformers, side-stepping the limitations led by the memory capacity of a single GPU.
Large language models, due to their significant size and long sequences, can place a considerable strain on graphical unit…
Large language models and multimodal foundation models like GPT4V, Claude, and Gemini, that blend visual encoders and language models, have made profound strides in the realms of Natural Language Processing (NLP) and Natural Language Generation (NLG). They show impressive performance when working with text-only inputs or a combination of image and text-based inputs. Nonetheless, queries…
Deep learning researchers have long been grappling with the challenge of designing a unifying framework for neural network architectures. Existing models are typically defined by a set of constraints or a series of operations they must execute. While both these approaches are beneficial, what's been lacking is a unified system that seamlessly integrates these two…
The development of large language models (LLMs) has historically been English-centric. While this has often proved successful, it has struggled to capture the richness and diversity of global languages. This issue is particularly pronounced with languages such as Korean, which boasts unique linguistic structures and deep cultural contexts. Nevertheless, the field of artificial intelligence (AI)…
Deep learning architectures require substantial resources due to their vast design space, lengthy prototyping periods, and high computational costs related to large-scale model training and evaluation. Traditionally, improvements in architecture have come from heuristic and individual experience-driven development processes, as opposed to systematic procedures. This is further complicated by the combinatorial explosion of possible designs…
Large Language Models (LLMs) have become increasingly influential in many fields due to their ability to generate sophisticated text and code. Trained on extensive text databases, these models can translate user requests into code snippets, design specific functions, and even create whole projects from scratch. They have numerous applications, including generating heuristic greedy algorithms for…
Recent advancements in large language models (LLMs) and Multimodal Foundation Models (MMFMs) have sparked a surge of interest in large multimodal models (LMMs). LLMs and MMFMs, including models such as GPT-4 and LLaVA, have demonstrated exceptional performance in vision-language tasks, including Visual Question Answering and image captioning. However, these models also require high computational resources,…
Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) are key advancements in artificial intelligence (AI) capable of generating text, interpreting images, and understanding complex multimodal inputs, mimicking human intelligence. However, concerns arise due to their potential misuse and vulnerabilities to jailbreak attacks, where malicious inputs trick the models into generating harmful or objectionable…
The Theory of Inventive Problem Solving (TRIZ) is a widely recognized method of ideation that uses the knowledge derived from a large, ongoing patent database to systematically invent and solve engineering problems. TRIZ is increasingly incorporating various aspects of machine learning and natural language processing to enhance its reasoning process.
Now, researchers from both the Singapore…
Artificial intelligence, particularly large language models (LLMs), faces the critical challenge of balancing model performance and practical constraints such as privacy, cost, and device compatibility. Large cloud-based models that offer high-accuracy rely on constant internet connectivity, raising potential issues of privacy breaches and high costs. Deploying these models on edge devices introduces further challenges in…
The transformer model has become a crucial technical component in AI, transforming areas such as language processing and machine translation. Despite its success, a common criticism is its standard method of uniformly assigning computational resources across an input sequence, failing to acknowledge the varying computational demands of different parts of a data sequence. This simplified…