Over 100 volunteers sampled meatballs made with the nutritious fruit. Breakthroughs, discoveries, and DIY tips sent every weekday. Finely ground, freeze-dried apple leftovers may become a sustainable secret ingredient in many meat dishes. In recent taste tests at Cornell University, more than 100 volunteers could barely tell the difference between 100-percent pure meatballs and alternatives featuring as much as 20 percent fruit waste. As the food researchers behind this culinary concoction explained in their study published in the, the supplemental additive may also help close a glaring gap in the food industry's circular loop.

In the rapidly evolving landscape of online recipe sharing within a globalized context, there has been a notable surge in research towards comprehending and generating food recipes. Recent advancements in large language models (LLMs) like GPT-2 and LLaVA have paved the way for Natural Language Processing (NLP) approaches to delve deeper into various facets of food-related tasks, encompassing ingredient recognition and comprehensive recipe generation. Despite impressive performance and multi-modal adaptability of LLMs, domain-specific training remains paramount for their effective application. This work evaluates existing LLMs for recipe generation and proposes LLaVA-Chef, a novel model trained on a curated dataset of diverse recipe prompts in a multi-stage approach. First, we refine the mapping of visual food image embeddings to the language space. Second, we adapt LLaVA to the food domain by fine-tuning it on relevant recipe data. Third, we utilize diverse prompts to enhance the model's recipe comprehension. Finally, we improve the linguistic quality of generated recipes by penalizing the model with a custom loss function. LLaVA-Chef demonstrates impressive improvements over pretrained LLMs and prior works. A detailed qualitative analysis reveals that LLaVA-Chef generates more detailed recipes with precise ingredient mentions, compared to existing approaches.

The AI community has embraced multi-sensory or multi-modal approaches to advance this generation of AI models to resemble expected intelligent understanding. Combining language and imagery represents a familiar method for specific tasks like image captioning or generation from descriptions. This paper compares these monolithic approaches to a lightweight and specialized method based on employing image models to label objects, then serially submitting this resulting object list to a large language model (LLM). This use of multiple Application Programming Interfaces (APIs) enables better than 95% mean average precision for correct object lists, which serve as input to the latest Open AI text generator (GPT-4). To demonstrate the API as a modular alternative, we solve the problem of a user taking a picture of ingredients available in a refrigerator, then generating novel recipe cards tailored to complex constraints on cost, preparation time, dietary restrictions, portion sizes, and multiple meal plans. The research concludes that monolithic multimodal models currently lack the coherent memory to maintain context and format for this task and that until recently, the language models like GPT-2/3 struggled to format similar problems without degenerating into repetitive or non-sensical combinations of ingredients. For the first time, an AI chef or cook seems not only possible but offers some enhanced capabilities to augment human recipe libraries in pragmatic ways. The work generates a 100-page recipe book featuring the thirty top ingredients using over 2000 refrigerator images as initializing lists.

Gary is implementing AI concepts from his Aerospace industry background onto use cases in retail and insurance. Sramana Mitra: Let's start by introducing you and Daisy Intelligence. Gary Saarenvirta: I'm the Founder and CEO of Daisy Intelligence. Daisy Intelligence is an AI platform. We help our clients make smarter operating decisions. Our mission is to empower human beings to do what humans are very good at by letting machines do what machines are good at. We have a couple of uses cases we address today in retail. It's a large segment of our business. We help retailers make smarter merchandise planning decisions. We help them to decide what products to promote, what prices to charge, and how much inventory to allocate. Our system delivers the decision. It's an autonomous decision-making system with no human in the loop. We deliver the answer to our clients. In the long run, our mission is to change the role of the human and let the machines do some of these beyond

I cannot put exact words on my frustration. I have been a Data Scientist for 3 years and I have been struggling for most of these 3 years to explain to my family what the purpose and day to day activities of my jobs are. First of all, my family only speaks French and Data Scientist does not translate fully to French. Officially, the French government proposed a translation: "expert en métadonnées". Spoiler Alert, it means nothing to anyone.