In this paper, we perform an exhaustive evaluation of different representations to address the intent classification problem in a Spoken Language Understanding (SLU) setup. We benchmark three types of systems to perform the SLU intent detection task: 1) text-based, 2) lattice-based, and a novel 3) multimodal approach. Our work provides a comprehensive analysis of what could be the achievable performance of different state-of-the-art SLU systems under different circumstances, e.g., automatically- vs. manually-generated transcripts. We evaluate the systems on the publicly available SLURP spoken language resource corpus. Our results indicate that using richer forms of Automatic Speech Recognition (ASR) outputs, namely word-consensus-networks, allows the SLU system to improve in comparison to the 1-best setup (5.5% relative improvement). However, crossmodal approaches, i.e., learning from acoustic and text embeddings, obtains performance similar to the oracle setup, a relative improvement of 17.8% over the 1-best configuration, being a recommended alternative to overcome the limitations of working with automatically generated transcripts.

Subgraph-wise sampling -- a promising class of mini-batch training techniques for graph neural networks (GNNs -- is critical for real-world applications. During the message passing (MP) in GNNs, subgraph-wise sampling methods discard messages outside the mini-batches in backward passes to avoid the well-known neighbor explosion problem, i.e., the exponentially increasing dependencies of nodes with the number of MP iterations. However, discarding messages may sacrifice the gradient estimation accuracy, posing significant challenges to their convergence analysis and convergence speeds. To address this challenge, we propose a novel subgraph-wise sampling method with a convergence guarantee, namely Local Message Compensation (LMC). To the best of our knowledge, LMC is the first subgraph-wise sampling method with provable convergence. The key idea is to retrieve the discarded messages in backward passes based on a message passing formulation of backward passes. By efficient and effective compensations for the discarded messages in both forward and backward passes, LMC computes accurate mini-batch gradients and thus accelerates convergence. Moreover, LMC is applicable to various MP-based GNN architectures, including convolutional GNNs (finite message passing iterations with different layers) and recurrent GNNs (infinite message passing iterations with a shared layer). Experiments on large-scale benchmarks demonstrate that LMC is significantly faster than state-of-the-art subgraph-wise sampling methods.

Learning to teach with ChatGPT is a new and somewhat controversial practice. Much of the conversation in education since the advanced AI chatbot debuted in late 2022 has focused on how to prevent ChatGPT cheating (opens in new tab). However, many educators now teach with ChatGPT and find it a helpful and versatile tool. Just remember ChatGPT is designed for use by adults 18 and up, and is still new technology with potential uses and misuses that are not yet fully understood. For example, technology related to ChatGPT that has been integrated with the Bing search engine for select users recently made headlines for saying disturbing things (opens in new tab) to a New York Times reporter.

Cheating in student assessments and exams has long been challenging for universities worldwide. Students are under increasing pressure to perform well, and some cheat to achieve good grades. This has led to universities implementing strict measures to prevent cheating, but new approaches are required with the advent of new technologies. One such technology that has the potential to help universities eliminate cheating is Chat GPT4. As universities worldwide continue to grapple with the challenge of eliminating cheating on student assessments and exams, the potential of new technologies like Chat GPT4 to help solve the problem is undeniable. However, some argue that such technologies may contribute to the problem by making cheating easier for students.

After the release of the powerful AI bot ChatGPT, its creators, Open AI, have released their latest AI model, which will be called GPT-4. The new model is only available to ChatGPT Plus subscribers for $20 monthly. The company has announced that the new addition to the AI family exhibits human-like performance on professional and economic tasks. It took the bar exams written by law school students before they could become professionals, and after the exams, GPT-4 was among the top 10% of test takers. GPT-4 has a lot of flexible capabilities, the biggest being that it is multimodal.

One of my professional goals this year was to learn more about artificial intelligence (AI). Over the course of the past year, there have been a lot of stories coming out about how schools are adding the concept of artificial intelligence into their curriculum or trying to weave it into different courses offered. The purpose is to help students better understand its capabilities and how it might impact the future of learning and the future of work. When I did some research earlier this year, I was amazed at some of the different uses of artificial intelligence that we interact with each day, and may not realize. A quick Google search of the term "artificial intelligence" turns up 518 million results in .17

Artificial intelligence (AI) has been a topic of discussion and research for decades. However, it is only in recent years that we have seen significant advancements in this field. AI is now being used in various industries, including healthcare, finance, and transportation, to name a few. With these advancements and applications, it is clear that AI is here to stay and will shape the future in various ways. So, what does the future of AI look like?

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. For the two weeks that the Hackneys' baby girl lay in a Pittsburgh hospital bed weak from dehydration, her parents rarely left her side, sometimes sleeping on the fold-out sofa in the room. They stayed with their daughter around the clock when she was moved to a rehab center to regain her strength. Finally, the 8-month-old stopped batting away her bottles and started putting on weight again. "She was doing well and we started to ask when can she go home," Lauren Hackney said.

Last week, WIRED published a series of in-depth, data-driven stories about a problematic algorithm the Dutch city of Rotterdam deployed with the aim of rooting out benefits fraud. In partnership with Lighthouse Reports, a European organization that specializes in investigative journalism, WIRED gained access to the inner workings of the algorithm under freedom-of-information laws and explored how it evaluates who is most likely to commit fraud. We found that the algorithm discriminates based on ethnicity and gender--unfairly giving women and minorities higher risk scores, which can lead to investigations that cause significant damage to claimants' personal lives. An interactive article digs into the guts of the algorithm, taking you through two hypothetical examples to show that while race and gender are not among the factors fed into the algorithm, other data, such as a person's Dutch language proficiency, can act as a proxy that enables discrimination. The project shows how algorithms designed to make governments more efficient--and which are often heralded as fairer and more data-driven--can covertly amplify societal biases.

"Turn Ideas into Python Programs with ChatGPT" is a unique online course that teaches you how to use ChatGPT, a powerful AI language model, to generate Python code without any prior experience in programming. This course is perfect for individuals who want to learn how to create Python automations and apps but have little to no coding experience. Instead of learning to code in Python, you'll learn how to write good ChatGPT queries that generate Python code. You'll then run this code in your computer and get the output which could be some generated files, web app, desktop GUI, a data analysis graph, etc. The course is divided into five sections, each of which focuses on a different aspect of using ChatGPT to build Python automations and apps.