We introduce Behavior from Language and Demonstration (BLADE), a framework for long-horizon robotic manipulation by integrating imitation learning and model-based planning. BLADE leverages language-annotated demonstrations, extracts abstract action knowledge from large language models (LLMs), and constructs a library of structured, high-level action representations. These representations include preconditions and effects grounded in visual perception for each high-level action, along with corresponding controllers implemented as neural network-based policies. BLADE can recover such structured representations automatically, without manually labeled states or symbolic definitions. BLADE shows significant capabilities in generalizing to novel situations, including novel initial states, external state perturbations, and novel goals. We validate the effectiveness of our approach both in simulation and on real robots with a diverse set of objects with articulated parts, partial observability, and geometric constraints.

We present a novel task and benchmark for evaluating the ability of text-to-image(T2I) generation models to produce images that fit commonsense in real life, which we call Commonsense-T2I. Given two adversarial text prompts containing an identical set of action words with minor differences, such as "a lightbulb without electricity" v.s. "a lightbulb with electricity", we evaluate whether T2I models can conduct visual-commonsense reasoning, e.g. produce images that fit "the lightbulb is unlit" vs. "the lightbulb is lit" correspondingly. Commonsense-T2I presents an adversarial challenge, providing pairwise text prompts along with expected outputs. The dataset is carefully hand-curated by experts and annotated with fine-grained labels, such as commonsense type and likelihood of the expected outputs, to assist analyzing model behavior. We benchmark a variety of state-of-the-art (sota) T2I models and surprisingly find that, there is still a large gap between image synthesis and real life photos--even the DALL-E 3 model could only achieve 48.92% on Commonsense-T2I, and the stable diffusion XL model only achieves 24.92% accuracy. Our experiments show that GPT-enriched prompts cannot solve this challenge, and we include a detailed analysis about possible reasons for such deficiency. We aim for Commonsense-T2I to serve as a high-quality evaluation benchmark for T2I commonsense checking, fostering advancements in real life image generation.

During an interview with the BBC last year, Google's senior vice-president for devices and services, Rick Osterloh, pondered whether a homeowner should disclose the presence of smart home devices to guests. "I would, and do, when someone enters into my home," he said. When your central heating thermostat asks for your phone number, your TV knows what you like to watch and hackers can install spyware in your home through a lightbulb security flaw, perhaps it's time we all started taking smart home privacy issues more seriously. Just this week the National Cyber Security Centre issued a warning to owners of smart cameras and baby monitors to review their security settings. You can get a quick overview of privacy options for many smart home devices using the Mozilla "*privacy not included" guide; however if you've already invested in particular technology, all is not lost.

Smart home technology has become increasingly popular in the latter half of the last decade and the 2020s could see it become even more mainstream. Many of the products are designed with the promise to make life easier and can help increase the security of a home with the likes of smart locks, doorbells and cameras becoming increasingly popular. Smart technology developers also claim it can help people save money. For example, British Gas says that its Hive smart thermostat could save users up to £120 a year on energy bills. Technology is expected to play an even bigger part in home security in 2020 with more residential properties equipping themselves with solutions that can be accessed remotely via mobile devices including smartphones, tablets and laptops.

Alexa to change the color of a lamp from white to red is about as cool as it comes. And gosh, you can make a similar request to the Google Assistant and go from red to blue, or orange. Siri will dim the lights for you and, like the other personal assistants, turn the lights on and off, on command. Is it any wonder that smart lighting has become on the most popular categories in the growing smart home space? Jiggling with your existing door lock and exchanging it for a smart model, or adding a new doorbell with a video camera can be a chore.

Voice assistants made popular by Amazon and Google are seemingly everywhere in the home - from internet-connected refrigerators, to toilets and lightbulbs. They bring with them the benefit of convenience as a growing number of users can now complete everyday tasks, like locking their door or turning on the light, with just their voice. But the always-on nature of internet-connected devices has raised some concerns over just how much data these applications are collecting and what they're doing with it, according to Bloomberg. A woman is seen controlling her Philips Hue smart lightbulb with her voice assistant. Concerns have grown around how much data these applications are collecting and how it's being used In the past, if users asked Alexa to turn on their smart bulb, Alexa would transmit code to the device to check if it was on or off, receive confirmation that it was off and then tell it to turn on, Bloomberg noted.

How it works: Control hot water and heating from your smartphone. How it works: A lightbulb you can dim, or turn on and off, using your phone. How it works: A smart speaker you control by speaking to Alexa, an intelligent'virtual assistant'. Functions include answering questions and playing music. How it works: Be alerted on your phone when someone rings your doorbell, and speak to that person via a video link, regardless of where you are.

How many AI people does it take to change a lightbulb? One to figure out how to describe lightbulb changing in first order logic. One to figure out how to describe lightbulb changing in second order logic. One to show the adequacy of FOL. One to show the inadequacy of FOL.

Echo Plus is one of the latest Alexa-enabled devices. You'd like to say "Alexa, turn off the lights," or "Alexa, lock the door," and have it all happen without you having to get up out of your lounge chair. No question, but it will cost you. Based on the costs of purchasing items, trying to install them and then giving up and then paying someone else for installation, USA TODAY estimates a figure of at least $2,200 to get started with smart lighting, doorbell, lock, thermostat and security. Compared to the options a few years ago, that's a bargain.

Machine Learning is a new term that re-packages well-established statistical techniques that have been around for decades. Machine Learning reacts to new information by learning. The "learning" happens by observing the past--sometimes the very recent past. Machine Learning can deliver on the promise of self-adjusting algorithms that can react to new information adeptly. However, before employing Machine Learning-driven predictions, ask this question: Is operating in a reactive mode our best option?