Fine-tuning techniques based on Large Pretrained Language Models (LPLMs) have been proven to significantly enhance model performance on a variety of downstream tasks and effectively control the output behaviors of LPLMs. Recent studies have proposed numerous methods for fine-tuning a small number of parameters based on open-source LPLMs, reducing the demand for computational and storage resources. Among these, reparameterization fine-tuning methods represented by LoRA (Low-Rank Adaptation) have gained popularity. We find that although these methods perform well in many aspects, there is still considerable room for improvement in terms of complex task adaptability, performance, stability, and algorithm complexity. In response to this, inspired by the idea that the functions of the brain are shaped by its geometric structure, this paper integrates this idea into LoRA technology and proposes a new matrix transformation-based reparameterization method for efficient fine-tuning, named Matrix-Transformation based Low-Rank Adaptation (MTLoRA). The spatiotemporal patterns of brain neural activity are the excitation of different wavelength characteristic patterns of its geometric structure. MTLoRA aims to dynamically alter its spatial geometric structure by applying a transformation-matrix T to perform linear transformations, such as rotation, scaling, and translation, on the task-specific parameter matrix, generating new matrix feature patterns (eigenvectors) to mimic the fundamental influence of complex geometric structure feature patterns in the brain on functions, thereby enhancing the model's performance in downstream tasks. The transformation-matrix T contains four different structures, each designed to simulate the geometric feature patterns of the brain at different levels. In Natural Language Understanding (NLU) tasks, it is evaluated using the GLUE benchmark test, and the results reveal that MTLoRA achieves an overall performance increase of about 1.0% across eight tasks and reduces the standard deviation by 0.7% in the Corpus of Linguistic Acceptability (CoLA) task; in Natural Language Generation (NLG) tasks, MTLoRA improves performance by an average of 0.95% and 0.56% in the DART and WebNLG tasks, respectively.

Language models (LMs) may lead their users to make suboptimal downstream decisions when they confidently hallucinate. This issue can be mitigated by having the LM verbally convey the probability that its claims are correct, but existing models cannot produce text with calibrated confidence statements. Through the lens of decision-making, we formalize linguistic calibration for long-form generations: an LM is linguistically calibrated if its generations enable its users to make calibrated probabilistic predictions. This definition enables a training framework where a supervised finetuning step bootstraps an LM to emit long-form generations with confidence statements such as "I estimate a 30% chance of..." or "I am certain that...", followed by a reinforcement learning step which rewards generations that enable a user to provide calibrated answers to related questions. We linguistically calibrate Llama 2 7B and find in automated and human evaluations of long-form generations that it is significantly more calibrated than strong finetuned factuality baselines with comparable accuracy. These findings generalize under distribution shift on question-answering and under a significant task shift to person biography generation. Our results demonstrate that long-form generations may be calibrated end-to-end by constructing an objective in the space of the predictions that users make in downstream decision-making.

OpenAI just announced that it recently conducted a small-scale preview of a new tool called Voice Engine. This is a voice cloning technology that can mimic any speaker by analyzing a 15-second audio sample. The company says it generates "natural-sounding speech" with "emotive and realistic voices." The technology is based on the company's pre-existing text-to-speech API and it has been in the works since 2022. OpenAI has already been using a version of the toolset to power the preset voices available in the current text-to-speech API and the Read Aloud feature. There are a bunch of samples on the company's official blog and they sound eerily close to the real thing.

Luke Bryan speaks during the signing of the ELVIS Act to Protect Voice & amp; Likeness in Age of AI event at Robert's Western World on March 21, 2024, in Nashville, Tennessee. 'AMAZING PRECEDENT': Luke Bryan is celebrating new protections from artificial intelligence for musicians in Nashville. Luke Bryan has high praise for the Tennessee state government over its new AI regulation law. ELECTION THREAT: Former Secretary of State Hillary Clinton described herself as a victim of election disinformation during a panel discussion on Thursday, and warned that the advancement of artificial intelligence (AI) will make her experience "look primitive." LEVEL UP: Google has developed an artificial intelligence system that can play video games like a human and take orders from players and could eventually even have real-world implications down the line.

Officially, the appointment started in January and lasts three months. But Reben's relationship with the San Francisco–based AI firm seems casual: "It's a little fuzzy, because I'm the first, and we're figuring stuff out. I'm probably going to keep working with them." In fact, Reben has been working with OpenAI for years already. Five years ago, he was invited to try out an early version of GPT-3 before it was released to the public. "I got to play around with that quite a bit and made a few artworks," he says.

Artificial intelligence startup OpenAI has been teasing its new AI video generator, Sora, on social media in recent weeks. Last week, it revealed that it had also given actors and directors in Hollywood a first look at the technology – and a chance to try it out – before Sora is launched publicly. OpenAI published a blog post on March 24 titled Sora's First Impressions, showcasing the work that several creative studios and directors had produced using the video generator. Some media experts speculate that Sora will be extremely disruptive to the film creative industry. Al Jazeera spoke to one executive who works in Hollywood, who asked us not to reveal his identity due to the sensitive nature of the subject.

Ever since the rollout of ChatGPT in November 2022, many people in science, business, and media have been obsessed with AI. A cursory look at my own published work during that period fingers me as among the guilty. My defense is that I share with those other obsessives a belief that large language models are the leading edge of an epochal transformation. Maybe I'm swimming in generative Kool-Aid, but I believe AI advances within our grasp will change not only the way we work, but the structure of businesses, and ultimately the course of humanity. Not everyone agrees, and in recent months there's been a backlash. AI has been oversold and overhyped, some experts now opine.

Elon Musk's answer to ChatGPT is getting an update to make it better at math, coding and more. Musk's xAI has launched Grok-1.5 to early testers with "improved capabilities and reasoning" and the ability to process longer contexts. Going by xAI's numbers, Grok-1.5 appears to be a large improvement over Grok-1. It shot up to 50.6 percent in the MATH benchmark, over double the previous score. It also climbed to 90 percent and 74.1 percent in GSM8K (math word problems) and HumanEval (coding), respectively, compared to 62.9 percent and 63.2 percent before.

Large language models (LLMs) have increased interest in vision language models (VLMs), which process image-text pairs as input. Studies investigating the visual understanding ability of VLMs have been proposed, but such studies are still preliminary because existing datasets do not permit a comprehensive evaluation of the fine-grained visual linguistic abilities of VLMs across multiple languages. To further explore the strengths of VLMs, such as GPT-4V \cite{openai2023GPT4}, we developed new datasets for the systematic and qualitative analysis of VLMs. Our contribution is four-fold: 1) we introduced nine vision-and-language (VL) tasks (including object recognition, image-text matching, and more) and constructed multilingual visual-text datasets in four languages: English, Japanese, Swahili, and Urdu through utilizing templates containing \textit{questions} and prompting GPT4-V to generate the \textit{answers} and the \textit{rationales}, 2) introduced a new VL task named \textit{unrelatedness}, 3) introduced rationales to enable human understanding of the VLM reasoning process, and 4) employed human evaluation to measure the suitability of proposed datasets for VL tasks. We show that VLMs can be fine-tuned on our datasets. Our work is the first to conduct such analyses in Swahili and Urdu. Also, it introduces \textit{rationales} in VL analysis, which played a vital role in the evaluation.

The burgeoning development of generative artificial intelligence (GenAI) and the widespread adoption of large language models (LLMs) in educational settings have sparked considerable debate regarding their efficacy and acceptability.Despite the potential benefits, the assimilation of these cutting-edge technologies among educators exhibits a broad spectrum of attitudes, from enthusiastic advocacy to profound skepticism.This study aims to dissect the underlying factors influencing educators' perceptions and acceptance of GenAI and LLMs.We conducted a survey among educators and analyzed the data through the frameworks of the Technology Acceptance Model (TAM) and Innovation Diffusion Theory (IDT). Our investigation reveals a strong positive correlation between the perceived usefulness of GenAI tools and their acceptance, underscoring the importance of demonstrating tangible benefits to educators. Additionally, the perceived ease of use emerged as a significant factor, though to a lesser extent, influencing acceptance. Our findings also show that the knowledge and acceptance of these tools is not uniform, suggesting that targeted strategies are required to address the specific needs and concerns of each adopter category to facilitate broader integration of AI tools.in education.