We’re building the future of language AI
Innovating Multilingual Interaction Ethical AI Safety-Critical Applications Guided by European Values
Why Choose Us
In a world where digital communication is pivotal, ELOQUENCE emerges as a beacon of innovation, reshaping the landscape of conversational AI. By harnessing advanced language models, ELOQUENCE empowers businesses and developers to create more natural, efficient, and meaningful interactions. Our technology is not just about facilitating conversations; it's about elevating them to new heights of clarity and impact.
Advanced Multilingual Support
Our multilingual support dismantles language barriers, enabling seamless communication across borders. ELOQUENCE's technology is designed for everyone, everywhere, making every interaction personal and inclusive.
Bias Mitigation and Ethical AI
We're committed to shaping a future where technology respects ethical, legal, and societal norms. Our models are meticulously crafted to minimize biases, ensuring fairness and transparency in every conversation.
Sustainable AI Solutions
Eco-Friendly AI: 50% Less Data, 50% Less Energy. We're devoted to shrinking our environmental footprint and addressing ethical and legal concerns, ensuring our technology's positive impact on society and the planet.
Innovating Interaction
ELOQUENCE is pioneering the development of advanced voice/chat bot technologies aimed at transforming unstructured dialogues into comprehensible, safe, and unbiased interactions. Our focus is on creating self-learning, adaptable models for diverse languages and use cases, emphasizing sustainability and European values, especially in safety-critical applications. By enhancing conversational agents with our research, including partnerships with key European enterprises and deploying in scenarios like emergency services and smart home assistants, ELOQUENCE aims to set new standards in AI, ensuring privacy, bias detection, and empathetic user engagement.
Our Vision, Your Future
Description goes here
Advancing SLU in Safety-Critical Applications
Advanced SLU technologies at TRL=5, validated in contact centers and smart home assistants for enhanced safety and efficiency.
Hybrid LLMs for Complex Dialogue Analysis
New LLMs facilitate contextual dialogue, handle complex queries, and ensure explainable decision-making with human involvement.
Unbiased Trustworthy Conversational Agents
New LLM architectures use innovative training algorithms to reduce bias and enhance reliability and trust for end-users.
Framework for AI Dialogue Assessment
ELOQUENCE tech in call centers and smart homes assessed with traditional metrics, new criteria for safety-critical cases.
Pilot Tests for Tech Validation
ELOQUENCE will run four pilots, including virtual agents for call centers and smart homes in safety-critical scenarios.
Legal, Ethical, and Societal Alignment
ELOQUENCE prioritizes European values, including privacy, non-discrimination, and legal and ethical robustness, in line with the upcoming AI Act.
Impact in Numbers
Leveraging the innovative edge of ELOQUENCE, we're redefining conversational AI with a focus on seamless, insightful, and inclusive digital interactions. Our impact resonates through solid numbers: from enhancing user experience across dozens of languages to deploying cutting-edge pilots that set new standards in AI ethics and sustainability. ELOQUENCE isn't just about technology; it's about crafting a future where every interaction is empowered by understanding, efficiency, and respect.
Showcase: Explore our Demos
Decentralized AI for Smarter Homes
We’re redefining smart home innovation by introducing decentralized AI training to enhance privacy in product development. This approach not only speeds up the development cycle but also enriches Telefonica’s smart home solutions, including AURA, your cognitive assistant, and our smart climate control, ensuring your home is not just connected but truly intelligent.
Elevating Smart Home Experience with Telefonica
Technological innovation centers on introducing hierarchical federated learning for LLMs and speech applications within telco infrastructures, enhancing privacy with differential layers, and enabling future decentralized developments with FL toolkits.
Bias Detection in Social Contexts
This pilot leverages open-source data to craft scenarios and benchmarks that test language models against social norms and detect biases in sentence construction. We’re designing multilingual interaction scenarios that reflect diverse social and cultural aspects, aiming to ensure AI-generated communication is socially sensitive and compliant. Through dialogues that delve into cultural, racial, and gender considerations, our goal is to analyze and develop linguistic resources that help in evaluating AI responses in varied contexts, including introducing perturbations to inputs to pinpoint triggers for relevant information.
Advancing Socially Aware AI Conversations with CNR
Technological innovation encompasses the development of multilingual linguistic resources for bias detection, emotion analysis modules for context-appropriate emotional responses, and retroactive feedback mechanisms to enhance bias awareness in LLMs.
Empathetic Virtual Agents through Retrieval-Augmented LLMs
In this pilot, we aim to assess the effectiveness of our proposed LLM-based virtual agent within the demanding customer service sector. Our objective is to develop a virtual agent trained on diverse industry-specific conversational datasets—from finance and insurance to utilities and dining—designed to adapt to new businesses seamlessly without needing additional adjustments. This virtual agent will be equipped to access and leverage relevant information from an enterprise’s structured or semi-structured documents to meet the caller’s needs.
Enhancing User Interaction with Omilia
Technological innovation is highlighted by retrieval-augmented LLMs that adapt to new enterprises seamlessly, learn to utilize tools like web services to achieve user goals, and blend enterprise-specific knowledge with pretrained insights for empathetic responses
Enhancing AI Supervision in Critical Support Call Centres
This pilot enhances emergency health call centers by introducing AI-based dialogue supervision to support junior medical staff in handling complex situations. It focuses on simulating real-world scenarios, such as assisting new parents, to improve response quality and efficiency in critical services.
Optimizing Emergency Response with UNS
Technological innovation includes a dialogue corpus to support AI recommendations for medical staff and TTS-based caller handling during wait times. Performance metrics include corpus size and key contact center indicators, enhancing medical staff efficiency.
Meet Our Partners
Led by Telefónica, ELOQUENCE is an elite alliance of 16 partners across 10 countries. We bring together top R&D centers, academic leaders, and conversational AI innovators, focusing on groundbreaking advancements in natural language processing, robotics, and cognitive systems. Our consortium, a blend of research excellence and SME innovation, is at the forefront of developing AI technologies for practical, sustainable applications. Committed to transforming the future of human-machine communication, ELOQUENCE embodies our dedication to creating impactful, socially responsible AI solutions.
Curious about large language model capabilities?
ELOQUENCE is your gateway to understanding context-sensitive and bias-managed conversational systems, specially designed for safety-critical environments. Sign up for our newsletter to stay informed!