Towards a Human-like Chatbot

guest · Jan 30, 2020

Google's AI-based chatbot potentially surpasses all other chatbots available today
January 30, 2020
https://www.neowin.net/news/googles...-surpasses-all-other-chatbots-available-today

Chatbots have improved considerably in recent years by leveraging the power of machine learning and neural networks. But even today, most of them work in specific scenarios and not a lot of them can confabulate over a wide range of topics. That might change with Google's Meena.

In the paper "Towards a Human-like Open-Domain Chatbot", Google claims that it has made a chatbot that is superior to all chatbots today. After training on 341GB of social-media conversations and sporting a 2.6 billion parameter end-to-end trained neural conversational model, "Meena can conduct conversations that are more sensible and specific than existing state-of-the-art chatbots.", the researchers wrote.

Impressively, Meena scored well with a 79% on the new metric. This score places Google's new chatbot between the venerable Mitsuko which scored 56% and a human conversing partner who score 86% on the same test.
Click to expand...

But Google will not be releasing Meena's demo to the general public domain before putting its safety and potential biases to the test.
Click to expand...

Google AI: Towards a Conversational Agent that Can Chat About…Anything

Paper: Towards a Human-like Open-Domain Chatbot (PDF - 0.98 MB): https://arxiv.org/pdf/2001.09977

guest · Feb 2, 2020

Just how big a deal is Google’s new Meena chatbot model?
February 1, 2020
https://venturebeat.com/2020/02/01/just-how-big-a-deal-is-googles-new-meena-chatbot-model/

Technology behemoths like Google and Facebook have got us used to, even fatigued by, their never-ending string of impressive announcements of progress in the AI field. Nevertheless, when Google announced that it has built a “conversational agent that can chat about… anything,” even the most jaded amongst us had to pay attention.

One of the biggest challenges for bots is to handle the infinite possible phrases that a user might say and respond appropriately.
Nevertheless, Meena is incredibly impressive and represents a fascinating attempt to solve the problem. In the next few paragraphs, I will summarize what Google did and how this might impact conversational AI in the days, months, and years to come. [...]
The bottom line
Meena is a fantastic contribution to the chatbot space. It is hard to capture the enormity of the task Google has achieved here. But we need to be careful about how we communicate the results of that research. Descriptions such as “the bot that can chat about anything” or “the best chatbot” are not necessarily useful.
Click to expand...

As more and more conversational AI solutions enter our daily lives, we need to focus on what is most valuable for us as humans. Meena moves us closer to that goal but doesn’t quite get us there yet.
Click to expand...

guest · Apr 29, 2020

Facebook open-sources Blender, a chatbot people say ‘feels more human’
April 29, 2020
https://venturebeat.com/2020/04/29/...r-a-chatbot-that-people-say-feels-more-human/

Facebook AI Research (FAIR), Facebook’s AI and machine learning division, today detailed work on a comprehensive AI chatbot framework called Blender. FAIR claims that Blender, which is available in open source on GitHub, is the largest-ever open-domain chatbot and outperforms existing approaches to generating dialogue while “feel[ing] more human,” according to human evaluators.

FAIR says Blender is the culmination of years of research to combine empathy, knowledge, and personality into one system. To this end, the underlying models — which benefit from improved decoding and skill blending techniques — contain up to 9.4 billion parameters (configuration variables that define skill on a given problem), or 3.6 times more than previous systems.
Blender promises to make interactions with conversational AI systems like Alexa, Siri, and Cortana more natural than before, whether in enterprise, industrial, or consumer-facing contexts.

That’s because they’re able to ask and answer a wide range of questions; display knowledge about specific topics; and express sentiments like empathy, seriousness, or playfulness as circumstances dictate.
Click to expand...

[...] Three models were considered by size, ranging from 90 million parameters to 2.7 billion parameters to 9.4 billion parameters.
Evaluations
Post-training, the researchers evaluated Blender’s performance by comparing it with Google’s latest Meena chatbot, a machine learning model with 2.6 billion parameters.
[...]
The pretrained and fine-tuned Blender models with 90 million parameters, 2.7 billion parameters, and 9.4 billion parameters are available on GitHub, along with a script for interacting with the bot (with safety filtering built in). All code for model evaluation and fine-tuning, including the data sets themselves, is available in ParAI.
Click to expand...

Engadget: Facebook releases its 'Blender' chatbot as an open-source project

However, size is only important up to a point. “One interesting thing we found is scale isn't everything,” Roller said. “For instance, our 2.7 billion parameter model, actually slightly outperforms the 9.4 billion parameter model. What seems to be really important is that you scale these things at least until a certain point, but then it becomes supercritical to imbue these things with specialized skills and behaviors [to further refine its performance].”

FAIR has focused on three specific behaviors -- the ability to display empathy, personality and knowledge -- to further humanize Blender’s responses.

So far, people who have interacted with Blender seem to prefer it to other open-source models. During a recent test between it and Meena using the ACUTE-Eval method, 67 percent of respondents found Blender to sound more human, and 75 percent would rather engage with Blender than with Meena for longer conversations.
Click to expand...

Facebook: A state-of-the-art open source chatbot

Facebook AI has built and open-sourced Blender, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators.

Today we’re releasing the complete model, code, and evaluation set-up, so that other AI researchers will be able to reproduce this work and continue to advance conversational AI research.

Read the paper here.
Get the code here.
Click to expand...

guest · Jun 3, 2020

Facebook creates shopping data sets to make more humanlike chatbots
June 3, 2020
https://venturebeat.com/2020/06/03/facebook-shopping-data-sets-humanlike-chatbots/

Facebook researchers this week introduced Situated Interactive MultiModal Conversations (SIMMC), a novel research direction aimed at training AI chatbots that take actions like showing an object and explaining what it’s made of in response to images, memories of previous interactions, and individual requests.

In a technical paper, they detail new data sets designed for this purpose containing around 13,000 human-to-human dialogs across two domains — furniture and fashion — along with several tasks framed as objective evaluation protocols.
Click to expand...

For example, given the prompt “I want to buy some chairs — show me brown ones and tell me about the materials,” the assistant might reply with an image of brown chairs and the text “How do you like these? They have a solid brown color with a foam fitting.”
Facebook says that it’ll publicly release the data, annotations, code, and models in the future.

The company says its goal is to one day combine its approaches into a system that can serve up product recommendations on the fly, matched to individual tastes and styles — a sort of hardware-free take on the recently discontinued Echo Look, Amazon’s AI-powered camera that told customers how their outfits looked and kept track of their wardrobe.
Click to expand...

Paper: Situated and Interactive Multimodal Conversations (PDF - 1.08 MB): https://arxiv.org/pdf/2006.01460

guest · Jun 10, 2020

‘Bot or Not?’ – a game to train us to spot chatbots faking it as humans
June 10, 2020
https://nakedsecurity.sophos.com/20...rain-us-to-spot-chatbots-faking-it-as-humans/

Who doesn’t know their mother’s maiden name?!
A bot that’s trying to convince you it’s human but which hasn’t been programmed to answer that question or improvise very convincingly, that’s who. Or, as I said when I finished playing a new online Turing Test game called Bot or Not, NAILED IT!!

Bot or Not is an online game that pits people against either bots or humans. It’s up to players to figure out which they’re engaging with in the 3-minute game, in which they’re forced to question not only whether their opponent is human but exactly how human they themselves are.
Click to expand...

In their explanation of Bot or Not’s purpose, the game’s creators point to a recent Gartner industry report that predicted that by 2020, the average person will engage in more conversations with bots than with their spouses.
Think about it: how often do you talk to voice assistants like Siri or OK Google? Chatbots have become seamlessly integrated into our lives [...]
Click to expand...

ZDNet: Bot or Not tests if you know you are talking to a human or machine

The tool aims to raise awareness about the risks unchecked chatbots pose for society -- from more powerful forms of surveillance to increased scams and exploitation. The central idea behind Bot Or Not, is an AI that mimics real humans.

The online game engages people in thinking critically about artificial agents that pretend to be human.
Click to expand...

guest · Dec 6, 2021

Amazon Lex brings automated call transcript review to chatbot building
December 1, 2021

Today, at AWS re:Invent in Las Vegas, the company announced Amazon Lex in preview, a new tool that simplifies the chatbot training and design process by bringing a level of automation to it.

“We are excited to announce the Amazon Lex automated chat bot builder, a new capability that reduces bot design from weeks to just hours,” Swami Sivasubramanian, VP of Amazon AI told the audience at the AI and machine learning keynote today.
It does this by taking advantage of advanced natural language understanding powered by deep learning techniques. In fact, he said that developers can now create a foundational chatbot designed using historical call transcripts in just a few clicks.
Click to expand...

When you think about a common use case for AI, chatbots certainly come to mind. They may be designed for in-house use to answer questions about how to order a new computer or get your newborn child on the company health insurance, or they may act as a customer service front end to collect vital information and answer simple questions, while funneling more complex questions to a human customer service agent.

Amazon Lex is available starting today in preview. Developers can use Lex for free during the preview phase, but will be charged based on the time it takes the tool to analyze a transcript and identify the intents once it is generally available.
Click to expand...

AWS Machine Learning Blog: Expedite conversation design with the automated chatbot designer in Amazon Lex

The automated chatbot designer expands the usability of Amazon Lex to the design phase. It uses machine learning (ML) to provide an initial bot design that you can then refine and launch conversational experiences faster. With the automated chatbot designer, Amazon Lex customers and partners get an easy and intuitive way of designing chatbots and can reduce bot design time from weeks to hours.

Organizations are rapidly adopting chatbots to increase self-service and improve customer experience at scale. Contact center chatbots automate common user queries and free up human agents to focus on solving more complex issues.
You can use Amazon Lex to build chatbots that deliver engaging user experiences and lifelike conversations.
Click to expand...

guest · Aug 5, 2022

Meta is putting its latest AI chatbot on the web for the public to talk to
What could possibly go wrong?
August 5, 2022

Meta’s AI research labs have created a new state-of-the-art chatbot and are letting members of the public talk to the system in order to collect feedback on its capabilities.

The bot is called BlenderBot 3 and can be accessed on the web. (Though, right now, it seems only residents in the US can do so.) BlenderBot 3 is able to engage in general chitchat, says Meta, but also answer the sort of queries you might ask a digital assistant, “from talking about healthy food recipes to finding child-friendly amenities in the city.”
Click to expand...

The bot is a prototype and built on Meta’s previous work with what are known as large language models or LLMS — powerful but flawed text-generation software of which OpenAI’s GPT-3 is the most widely known example.

A big feature of the chatbot is that it’s capable of searching the internet in order to talk about specific topics. Even more importantly, users can then click on its responses to see where it got its information from. BlenderBot 3, in other words, can cite its sources.
Click to expand...

By releasing the chatbot to the general public, Meta wants to collect feedback on the various problems facing large language models. Users who chat with BlenderBot will be able to flag any suspect responses from the system, and Meta says it’s worked hard to “minimize the bots’ use of vulgar language, slurs, and culturally insensitive comments.”
Click to expand...

The real prize is building a system that can conduct a conversation as free-ranging and natural as a human’s, and Meta says the only way to achieve this is to let bots have free-ranging and natural conversations.

“This lack of tolerance for bots saying unhelpful things, in the broad sense of it, is unfortunate,” says Williamson. “And what we’re trying to do is release this very responsibly and push the research forward.”

In addition to putting BlenderBot 3 on the web, Meta is also publishing the underlying code, training dataset, and smaller model variants. Researchers can request access to the largest model, which has 175 billion parameters, through a form here.
Click to expand...

Log in or Sign up

Towards a Human-like Chatbot

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

Log in or Sign up

Towards a Human-like Chatbot

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

guest Guest

Useful Searches