AI Beats Humans In Creativity, New Model For Deepfake Detection, AI For Emotional Support, & More!
Welcome to the AI Search newsletter. Here are the top updates in AI this week.
This AI Animation Tool CHANGES EVERYTHING
By inputting a single image and audio, such as singing or speaking, the method can generate videos with dynamic facial expressions, head poses, and various durations. The process involves two main stages: Frames Encoding, where features are extracted from the reference image and motion frames, and Diffusion Process, where audio embedding is processed and various attention mechanisms are employed to preserve character identity and modulate movements. The method supports different languages, portrait styles, and paces of audio, allowing for the creation of lifelike animated avatars in various settings, including cross-actor performances in different languages and styles.
Predicting Household Energy Costs Using AI and Google Street View
Using AI and Google Street View, researchers are predicting household energy costs to help address the high energy burden faced by low-income households in the United States. By analyzing passive design characteristics and demographic data, they have developed a model that can accurately predict energy expenses in the Chicago metropolitan area. These insights can be valuable for policymakers and urban planners in creating smart and sustainable cities, especially in vulnerable neighborhoods at risk of health hazards due to lack of affordable heating and cooling.
Humanoid Robot Company Gets Funding from OpenAI, Bezos, Nvidia
OpenAI teams up with Figure AI, a robotics startup based in Sunnyvale, California, to integrate AI systems into humanoid robots. The partnership, announced with $675 million in venture capital funding from Jeff Bezos, Microsoft, Nvidia, and others, aims to develop robots capable of performing tasks that humans may not want to do. This collaboration will involve building specialized AI models for Figure's robots, leveraging technologies like GPT language models, DALL-E image generation, and Sora video generation. The ultimate goal is to enhance the commercial prospects of Figure's humanoid robots by enabling them to process and reason from language.
Murf AI
Introducing Murf, the most versatile AI text-to-speech generator. Create studio-quality voice overs in minutes using lifelike AI voices suitable for podcasts, videos, presentations, and more. Choose from 120+ text to speech voices in 20+ languages, with the ability to add video, music, or images and sync them to the voiceover.
New Research For Detecting AI Deepfakes
In the ongoing fight against deepfake technology, researchers are working to develop more accurate detection methods. Recently, a novel approach has emerged, combining machine learning models to identify manipulated media content with unprecedented precision. By combining the miniXception CNN architecture with LSTM technology, they achieved a detection accuracy of 99.05% on the FaceSwap dataset, surpassing previous methods. With cross-dataset training, transfer learning, and focal loss strategies, this innovative approach offers hope in the ongoing battle against digital deception and the protection of public discourse and democracy.
AI Outperforms Humans In Tests Of Creativity
AI surpasses humans in tests of creative potential, as demonstrated by a recent study involving 151 participants competing against ChatGPT-4 in assessments of divergent thinking. Divergent thinking, represented by the ability to generate unique solutions to open-ended questions, was more effectively demonstrated by GPT-4 compared to the human participants. Despite this, the study authors note that AI lacks agency and depends on human interaction to prompt creative potential. While the AI produced more original and elaborate responses in the tests, human participants may still hold an advantage in real-world applicability and creative achievements. The study raises questions about the measures of creativity, asserting that AI's advancements do not necessarily equate to a threat to human creativity.
Luminar NEO
Luminar Neo is a new AI-driven creative image editor developed by Skylum Software. It’s designed to make complex editing quick and easy for all levels of photographers, from beginners to pros. Luminar Neo leverages artificial intelligence and 3D depth mapping, offering innovative tools such as the new Portrait Background Removal AI, Mask AI, and the Relight AI tool with 3D Depth Mapping. It can function as a standalone application for macOS and Windows, and can also integrate with Lightroom Classic, Photoshop, Photos for macOS, and Microsoft Photos as a plugin and extension. With Luminar Neo, you can transform your photos into the images you imagined, making creative image editing accessible and fun.
This New AI Offers Emotional Support
EmoAda is a new AI system developed by researchers that provides emotional support through chat. The system utilizes various forms of sensory data, such as voice, video, and text, to analyze a user's emotions and deliver personalized emotional support dialogues either in text form or through a digital avatar. EmoAda also suggests activities like guided meditation practices and music for relaxation based on the user's needs and difficulties mentioned. Initial test trials have shown that users appreciate the anonymity offered by EmoAda, allowing them to freely express their feelings and concerns without fear of judgment. This innovative AI system could potentially serve as a basic support service for individuals lacking access to professional psychological care and inspire the development of similar mental health-related digital platforms in the future.
Insect Mimicking Robot For Motion Detection
A research team at KAIST created an intelligent sensor that mimics the optic nerve of insects, offering high efficiency and speed while consuming minimal power. By utilizing memristor devices, the sensor can accurately predict motion, leading to potential applications in transportation, safety, and security systems. This breakthrough in neuromorphic computing signifies a significant advancement in the field of AI technology, with the device demonstrating the ability to predict vehicle paths with improved accuracy and energy efficiency. The simple structure of the sensor, consisting of two types of memristors and a resistor, enables the direct mimicry of insect visual intelligence pathways, showcasing its potential for integration into various innovative technologies.