Text-conditional diffusion models are able to generate high-fidelity images with diverse contents.However, linguistic representations frequently exhibit ambiguous descriptions of the envisioned objective imagery, requiring the incorporation of additional control signals to bolster the efficacy of text-guided diffusion models. In this work, we propose Cocktail, a pipeline to mix various modalities into one embedding, amalgamated with a generalized ControlNet (gControlNet), a controllable normalisation (ControlNorm), and a spatial guidance sampling method, to actualize multi-modal and spatially-refined control for text-conditional diffusion models. Specifically, we introduce a hyper-network gControlNet, dedicated to the alignment and infusion of the control signals from disparate modalities into the pre-trained diffusion model.

Automatic cocktail machines are silly, but also kind of fun. Here's how to choose between the Bartesian and Barsys devices. The machine on my kitchen table is a holy device, if your definition of "holy" is that it looks like a glowing halo and it's filled with spirits. The machine has taken up a task I consider sacred: making me a cocktail. In advance of holiday party season, I have been testing a pair of devices that promise an indulgent future, a life where machines can make you a passable Old Fashioned.

Recommender systems (RS) are currently being studied to mitigate limitations during cold-start conditions by leveraging modality information or introducing Agent concepts based on the exceptional reasoning capabilities of Large Language Models (LLMs). Meanwhile, food and beverage recommender systems have traditionally used knowledge graph and ontology concepts due to the domain's unique data attributes and relationship characteristics. On this background, we propose MARC, a multimodal and multi-task cocktail recommender system based on Agentic Retrieval-Augmented Generation (RAG) utilizing graph database under cold-start conditions. The proposed system generates high-quality, contextually appropriate answers through two core processes: a task recognition router and a reflection process. The graph database was constructed by processing cocktail data from Kaggle, and its effectiveness was evaluated using 200 manually crafted questions. The evaluation used both LLM-as-a-judge and human evaluation to demonstrate that answers generated via the graph database outperformed those from a simple vector database in terms of quality. The code is available at https://github.com/diddbwls/cocktail_rec_agentrag

A convent turned hotel, Caribou Coffee, and progressive coworking space called The Coven--plus more things to see and do while on a business trip to Minneapolis and St. Paul. All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links. Minnesota is the birthplace of the supercomputer, developed for code cracking during World War II. Tech giants of their day, including Cray Research and Control Data Corporation, were based in the Twin Cities.

-- Learning by Asking (LBA) enables robots to identify knowledge gaps during task execution and acquire the missing information by asking targeted questions. However, different tasks often require different types of questions, and how to adapt questioning strategies accordingly remains under-explored. This paper investigates human questioning behavior in two representative household service tasks: a Goal-Oriented task (refrigerator organization) and a Process-Oriented task (cocktail mixing). Through a human-human study involving 28 participants, we analyze the questions asked using a structured framework that encodes each question along three dimensions: acquired knowledge, cognitive process, and question form. Our results reveal that participants adapt both question types and their temporal ordering based on task structure. Goal-Oriented tasks elicited early inquiries about user preferences, while Process-Oriented tasks led to ongoing, parallel questioning of procedural steps and preferences. These findings offer actionable insights for developing task-sensitive questioning strategies in LBA-enabled robots for more effective and personalized human-robot collaboration. Active learning has become an increasingly influential paradigm in robotics, enabling robots to iteratively query human users (oracles) for labels on informative samples during human-robot interaction. This process reduces uncertainty by enabling the robot to selectively acquire information about ambiguous or unfamiliar situations through human input [1].

--Recently, large language models (LLMs) have been able to handle longer and longer contexts. However, a context that is too long may cause intolerant inference latency and GPU memory usage. Existing methods propose mixed-precision quantization to the key-value (KV) cache in LLMs based on token granularity, which is time-consuming in the search process and hardware inefficient during computation. This paper introduces a novel approach called Cocktail, which employs chunk-adaptive mixed-precision quantization to optimize the KV cache. Cocktail consists of two modules: chunk-level quantization search and chunk-level KV cache computation. Chunk-level quantization search determines the optimal bitwidth configuration of the KV cache chunks quickly based on the similarity scores between the corresponding context chunks and the query, maintaining the model accuracy. Furthermore, chunk-level KV cache computation reorders the KV cache chunks before quantization, avoiding the hardware inefficiency caused by mixed-precision quantization in inference computation. Extensive experiments demonstrate that Cocktail outperforms state-of-the-art KV cache quantization methods on various models and datasets.

During the early stages of interface design, designers need to produce multiple sketches to explore a design space. Design tools often fail to support this critical stage, because they insist on specifying more details than necessary. Although recent advances in generative AI have raised hopes of solving this issue, in practice they fail because expressing loose ideas in a prompt is impractical. In this paper, we propose a diffusion-based approach to the low-effort generation of interface sketches. It breaks new ground by allowing flexible control of the generation process via three types of inputs: A) prompts, B) wireframes, and C) visual flows. The designer can provide any combination of these as input at any level of detail, and will get a diverse gallery of low-fidelity solutions in response. The unique benefit is that large design spaces can be explored rapidly with very little effort in input-specification. We present qualitative results for various combinations of input specifications. Additionally, we demonstrate that our model aligns more accurately with these specifications than other models.

Tabular datasets are the last "unconquered castle" for deep learning, with traditional ML methods like Gradient-Boosted Decision Trees still performing strongly even against recent specialized neural architectures. In this paper, we hypothesize that the key to boosting the performance of neural networks lies in rethinking the joint and simultaneous application of a large set of modern regularization techniques. As a result, we propose regularizing plain Multilayer Perceptron (MLP) networks by searching for the optimal combination/cocktail of 13 regularization techniques for each dataset using a joint optimization over the decision on which regularizers to apply and their subsidiary hyperparameters. We empirically assess the impact of these regularization cocktails for MLPs in a large-scale empirical study comprising 40 tabular datasets and demonstrate that (i) well-regularized plain MLPs significantly outperform recent state-of-the-art specialized neural network architectures, and (ii) they even outperform strong traditional ML methods, such as XGBoost.