Luke Mahoney (MLabs)

Project owner

NEURAL SEARCH

Neural Search is a new way to incorporate Monte Carlo Tree Search into neural networks. Unlike similar techniques, Neural Search builds the exploration and optimization of the search into the layers of the network. The search itself is learnable, and provides us with a means to develop guided search algorithms trained on data. We have demonstrated that the algorithm can learn robust solutions to search problems. We propose further development of the approach to large scale problems, and to demonstrate the effectiveness of the algorithm by training it to play Go, a standard achieved AlphaGo.

Total for the project is $150K over 11 months, and it will be resourced by 3 senior MLabs members.

Sam Gittis

Project owner

AI Song Starter: Innovate Music Creation

AI Song Starter is a revolutionary tool designed to streamline music production by generating custom loops and stems based on text prompts or audio inputs. This tool helps musicians overcome writer's block, save time, and enhance creativity. By integrating with SingularityNET AI Marketplace, it democratizes access to advanced music production tools, fostering creativity and innovation across the music industry. This initiative supports over 60 million musicians worldwide, driving broader adoption and engagement within the SingularityNET ecosystem.

Adam Vandervorst

Project owner

Scalable MeTTa Knowledge Graphs

Use the new performant Hyperon backend, MeTTa Optimal Reduction Kernel (MORK), to host knowledge graphs core to the SingularityNET ecosystem. By using Hyperon for this task, we have the opportunity to practice dogfooding on core capabilities like importing relevant data from existing platforms and file formats, storing large knowledge bases in a compressed yet accessible form, supporting both SPARQL and MeTTa queries, and communicating to other services. Furthermore, by the nature of Hyperon's Metagraphs, we can store the code, documentation, version control, and knowledge graph data in the same representation. All functionality would be available via a rich online MeTTa editor.

photrek

Project owner

Intelligent Ocular Image Generation

AI applications developed for detection of medical conditions are hindered by lack of available data for training. Photrek contributes to the goals of DF4 with this Miscellaneous Pool proposal by generating synthetic image libraries for diseases of the human vision system and making these capabilities available to the community. Photrek has built a team across three continents (Africa, Asia, North America) at four universities (U of Buea Cameroon, VIT India, KNUST Ghana, and SUNY Poly USA) with a team focused on image processing techniques applied to medical applications in a low cost environment within SNET goals for creation of democratic, decentralized, beneficial AGI.

photrek

Project owner

Reputation & Voting Weight System

This RFP design aims to develop a sophisticated system to enhance the SingularityNET ecosystem's reputation and voting mechanisms. It includes four key milestones: conducting a community workshop to refine system architecture, defining high-level requirements for a suite of microservices to export crucial data, developing an adaptive algorithm layer for dynamic engagement and voting score calculations, and constructing a robust knowledge graph for comprehensive data integration. These components will ensure scalability, transparency, and fairness in community-driven decision-making processes.

Devbasrahtop

Project owner

Next Trend LAB: Exploring Decentralized AI for IoT

Next Trend LAB aims to revolutionize SingularityNET by researching the integration of decentralized AI algorithms with IoT. This exploration promises to enhance automation, predictive maintenance, and decision-making in smart environments. By bridging AI with IoT, we'll unlock innovative solutions, driving the adoption of AGIX tokens and expanding SingularityNET's ecosystem.

Robin Spottiswoode

Project owner

Bridging SNET AI Services To Unity Apps

Our project aims to bridge the SingularityNET AI Marketplace with the Unity ecosystem, enabling seamless integration of AI services into 3D/VR/AR applications. We will develop Unity scripts and tools to simplify the discovery and interaction with AI services, specifically integrating the ‘Song Splitter’ and ‘AI Synthesizer’ into our 3D Music Studio app. This will unlock vast potential in gaming, education, and digital entertainment, providing over 2 million Unity developers access to powerful AI tools, fostering innovation, and expanding the SingularityNET ecosystem.

Almalgo_Labs

Project owner

Zero-Knowledge Secure LLM

We plan to use homomorphic encryption (HE) to enable zero-knowledge security in Large Language Models (LLMs) on SingularityNet's decentralized AI platform. It involves modifying LLMs to process encrypted data, and maintaining confidentiality throughout. The system uses distributed computing, optimizing performance via hardware accelerations and batching. Security is reinforced with regular audits and zero-knowledge proofs to meet global standards like GDPR, ensuring scalable, secure AI applications.

SingularityNet Marketplace ChatGPT plugin

Arbiteria proposes creating a ChatGPT plugin to address the lack of information about decentralized AI and the SingularityNet platform. This plugin, funded with $39,000, will provide ChatGPTPlus users with responses related to decentralized AI queries, including information about SingularityNet AI Marketplace listings and the benefits of decentralized AI.

The project aims to raise awareness among both technical and non-technical users in the decentralized AI space. It also serves as a discreet marketing asset, potentially increasing user traffic to the SingularityNet platform and transactions in the SNET marketplace.

The project is divided into four milestones, including building the plugin, incorporating SNET marketplace apps, surfacing specific SNET information, and providing public Beta support. Mitigation strategies are in place in case key stakeholders do not reach a consensus on the information to surface.

The project is open source and may involve voluntary revenue sharing with SNET/Deep Funding for larger or high-risk projects.

Anthony Oliko

Project owner

Phase 1 of Huggingface datasets library to SingularityNET Pipeline

Phase 1: Building the Hugging Face Datasets library to SingularityNET Pipeline Framework

In this phase, we propose to build the foundational framework for the Hugging Face Datasets library to the SingularityNET Pipeline. This pipeline will facilitate seamless integration and accessibility of Knowledge Graph Question Answer datasets via the Hugging Face Datasets library, providing an efficient interface for users to access and leverage these datasets for future AGI Integrations. The framework will serve as the backbone for the subsequent integration of specific datasets in Phase 2.

NIM Framework | Framework for integrating AI services in Singularity Marketplace and bots in Matrix | Matrix Messenger

NIM (Native Intelligence Messenger) seeks to streamline communication by integrating AI services directly into messaging platforms. NIM's framework, compatible with the decentralized Matrix protocol, enables users to seamlessly access various AI bots within a single messenger app. This revolutionary approach eliminates the need for users to switch between applications, enhancing efficiency and productivity. Key features include a unified, user-friendly interface, personalized interactions based on user history, natural and interactive dialogues with AI bots, stringent data security measures, continuous product development, and interoperability for creating unique AI service pipelines. With a budget of $180-210k, NIM plans to achieve these milestones and make AI integration into messaging a game-changer. Additionally, they seek input from AI experts, marketers, and product analysts and aim to hire a frontend developer to expedite development. The framework will be open source, while the messenger itself will be a closed team project.

Juana Attieh

Project owner

Simulating a Risky World

Photrek plans to develop Simulating Risky Worlds, a service for simulating virtual worlds with controllable uncertainty and relative risk levels. This service combines two components: Photrek's Risk-Aware Data Generator and Deep Brain's Open-Source World Models. They've secured $49,480 in funding to address the issue of brittle AI/ML algorithms, making them more robust to deviations from training data.

The solution utilizes Coupled Variational Autoencoder (CVAE) and includes a Memory Module with a Recurrent Neural Network (RNN) for predictive simulations. They aim to demonstrate the utility of Simulating a Risky World to potential customers, including organizations like Vantage Risk and NOAA Severe Storm Laboratory.

The project will be divided into milestones, including integrating CVAE with Dynamics and onboarding Simulating a Risky World. Hosting costs will cover Google Cloud servers, Ethereum transaction fees, and AGIX fees.

Photrek intends to offer 2.5% of revenue when it exceeds $5,000 as support to SingularityNET. Their software will be open source, with various licensing models for different components.

Memory-augmented LLMs: Retrieving Information using a Gated Encoder LLM (RIGEL)

MLabs, an AI and Blockchain consultancy, seeks funding through RFP3 - Memory-augmented LLMs. They propose addressing the shortcomings of large language models (LLMs), such as ChatGPT, concerning the veracity of generated responses. MLabs aims to develop a new module for LLMs, enabling them to reference the source material, providing explanation and evidence to support responses, thereby reducing the risk of factually incorrect outputs or hallucinations. Their solution involves compressing context vectors and storing them externally, using hierarchical self-attention for efficient data compression and retrieval.

The project includes three parts: large-scale contextual compression experiments, training and evaluating the compression and retrieval module, and integration with the existing LLM. The total funding request is USD 140,000 for a 36-week project. Risks include the novelty of the concept, vector retrieval system performance, machine learning training, software development, and project management, which they aim to mitigate with their experienced team and contingency plans.

MLabs has a proven track record in ML/AI deployments and aims to leverage their expertise to enhance LLMs.

Juana Attieh

Project owner

Architectural Design for Uncertain Knowledge Graphs

Photrek's proposal for the "Tools for Knowledge Graphs and LLMs Integration" project outlines their objective to address uncertainty in knowledge graphs (KGs) and their integration with large language models (LLMs). They aim to embed a model of uncertainty into KGs at both micro and macro levels. At the micro level, Photrek will introduce confidence measures for each fact in KGs, enhancing traditional probability-based models with upper and lower bounds based on relative risk aversion. This approach allows for a wider range of uncertainty computations in KG inference.

At the macro level, they will design a methodology for managing short-term and long-term memory in KGs using generative AI methods. This involves monitoring and updating short-term and long-term memory to facilitate effective KG inference while avoiding overconfidence and errors.

Photrek plans to collaborate closely with SingularityNET's KG and LLM teams to determine priorities, with the higher priority item resulting in a detailed architectural design documented in a research report. The ultimate goal is to improve fact-checking and knowledge graph updates.

Ammar Khairi

Project owner

Customizable Code Assistant

This project addresses the accessibility gap in using and developing large language models (LLMs) by making them more usable and reducing computational costs. The aim is to democratize LLMs' practicality by exploring transfer learning on small pre-trained models for new programming languages and tasks. The solution entails generating datasets, conducting performance-cost analysis, fine-tuning and evaluating models, creating GPU inference environments, and deploying a modular pipeline for customizable Code Assistants. The budget of $30,000 covers these milestones. Risks are mitigated through privacy measures, licensing compliance, and adapting to computation costs. The project's inspiration stems from previous collaborative work and aims to contribute to AI's democratization through accessible and open-source models and datasets.