-
Decentralized Technologies for AI Hubs
Authors:
Richard Blythman,
Mohamed Arshath,
Salvatore Vivona,
Jakub Smékal,
Hithesh Shaji
Abstract:
AI requires heavy amounts of storage and compute with assets that are commonly stored in AI Hubs. AI Hubs have contributed significantly to the democratization of AI. However, existing implementations are associated with certain benefits and limitations that stem from the underlying infrastructure and governance systems with which they are built. These limitations include high costs, lack of monet…
▽ More
AI requires heavy amounts of storage and compute with assets that are commonly stored in AI Hubs. AI Hubs have contributed significantly to the democratization of AI. However, existing implementations are associated with certain benefits and limitations that stem from the underlying infrastructure and governance systems with which they are built. These limitations include high costs, lack of monetization and reward, lack of control and difficulty of reproducibility. In the current work, we explore the potential of decentralized technologies - such as Web3 wallets, peer-to-peer marketplaces, storage and compute, and DAOs - to address some of these issues. We suggest that these infrastructural components can be used in combination in the design and construction of decentralized AI Hubs.
△ Less
Submitted 7 June, 2023;
originally announced June 2023.
-
A Blockchain Protocol for Human-in-the-Loop AI
Authors:
Nassim Dehouche,
Richard Blythman
Abstract:
Intelligent human inputs are required both in the training and operation of AI systems, and within the governance of blockchain systems and decentralized autonomous organizations (DAOs). This paper presents a formal definition of Human Intelligence Primitives (HIPs), and describes the design and implementation of an Ethereum protocol for their on-chain collection, modeling, and integration in mach…
▽ More
Intelligent human inputs are required both in the training and operation of AI systems, and within the governance of blockchain systems and decentralized autonomous organizations (DAOs). This paper presents a formal definition of Human Intelligence Primitives (HIPs), and describes the design and implementation of an Ethereum protocol for their on-chain collection, modeling, and integration in machine learning workflows.
△ Less
Submitted 19 November, 2022;
originally announced November 2022.
-
Is Decentralized AI Safer?
Authors:
Casey Clifton,
Richard Blythman,
Kartika Tulusan
Abstract:
Artificial Intelligence (AI) has the potential to significantly benefit or harm humanity. At present, a few for-profit companies largely control the development and use of this technology, and therefore determine its outcomes. In an effort to diversify and democratize work on AI, various groups are building open AI systems, investigating their risks, and discussing their ethics. In this paper, we…
▽ More
Artificial Intelligence (AI) has the potential to significantly benefit or harm humanity. At present, a few for-profit companies largely control the development and use of this technology, and therefore determine its outcomes. In an effort to diversify and democratize work on AI, various groups are building open AI systems, investigating their risks, and discussing their ethics. In this paper, we demonstrate how blockchain technology can facilitate and formalize these efforts. Concretely, we analyze multiple use-cases for blockchain in AI research and development, including decentralized governance, the creation of immutable audit trails, and access to more diverse and representative datasets. We argue that decentralizing AI can help mitigate AI risks and ethical concerns, while also introducing new issues that should be considered in future work.
△ Less
Submitted 3 November, 2022;
originally announced November 2022.
-
Libraries, Integrations and Hubs for Decentralized AI using IPFS
Authors:
Richard Blythman,
Mohamed Arshath,
Jakub Smékal,
Hithesh Shaji,
Salvatore Vivona,
Tyrone Dunmore
Abstract:
AI requires heavy amounts of storage and compute. As a result, AI developers are regular users of centralised cloud services such as AWS, GCP and Azure, compute environments such as Jupyter and Colab notebooks, and AI Hubs such as HuggingFace and ActiveLoop. There services are associated with certain benefits and limitations that stem from the underlying infrastructure and governance systems with…
▽ More
AI requires heavy amounts of storage and compute. As a result, AI developers are regular users of centralised cloud services such as AWS, GCP and Azure, compute environments such as Jupyter and Colab notebooks, and AI Hubs such as HuggingFace and ActiveLoop. There services are associated with certain benefits and limitations that stem from the underlying infrastructure and governance systems with which they are built. These limitations include high costs, lack of monetization and reward, lack of control and difficulty of reproducibility. At the same time, there are few libraries that allow data scientists to interact with decentralised storage in the language that data scientists are used to, and few hubs where they can discover and interact with AI assets. In this report, we explore the potential of decentralized technologies - such as Web3 wallets, peer-to-peer marketplaces, decentralized storage (IPFS and Filecoin) and compute, and DAOs - to address some of the above limitations. We showcase some of the libraries and integrations that we have built to tackle these issues, as well as a proof of concept of a decentralized AI Hub app, that all use IPFS as a core infrastructural component.
△ Less
Submitted 29 October, 2022;
originally announced October 2022.
-
Addressing Bias in Face Detectors using Decentralised Data collection with incentives
Authors:
M. R. Ahan,
Robin Lehmann,
Richard Blythman
Abstract:
Recent developments in machine learning have shown that successful models do not rely only on huge amounts of data but the right kind of data. We show in this paper how this data-centric approach can be facilitated in a decentralized manner to enable efficient data collection for algorithms. Face detectors are a class of models that suffer heavily from bias issues as they have to work on a large v…
▽ More
Recent developments in machine learning have shown that successful models do not rely only on huge amounts of data but the right kind of data. We show in this paper how this data-centric approach can be facilitated in a decentralized manner to enable efficient data collection for algorithms. Face detectors are a class of models that suffer heavily from bias issues as they have to work on a large variety of different data. We also propose a face detection and anonymization approach using a hybrid MultiTask Cascaded CNN with FaceNet Embeddings to benchmark multiple datasets to describe and evaluate the bias in the models towards different ethnicities, gender, and age groups along with ways to enrich fairness in a decentralized system of data labeling, correction, and verification by users to create a robust pipeline for model retraining.
△ Less
Submitted 28 October, 2022;
originally announced October 2022.
-
Jointformer: Single-Frame Lifting Transformer with Error Prediction and Refinement for 3D Human Pose Estimation
Authors:
Sebastian Lutz,
Richard Blythman,
Koustav Ghosal,
Matthew Moynihan,
Ciaran Simms,
Aljosa Smolic
Abstract:
Monocular 3D human pose estimation technologies have the potential to greatly increase the availability of human movement data. The best-performing models for single-image 2D-3D lifting use graph convolutional networks (GCNs) that typically require some manual input to define the relationships between different body joints. We propose a novel transformer-based approach that uses the more generalis…
▽ More
Monocular 3D human pose estimation technologies have the potential to greatly increase the availability of human movement data. The best-performing models for single-image 2D-3D lifting use graph convolutional networks (GCNs) that typically require some manual input to define the relationships between different body joints. We propose a novel transformer-based approach that uses the more generalised self-attention mechanism to learn these relationships within a sequence of tokens representing joints. We find that the use of intermediate supervision, as well as residual connections between the stacked encoders benefits performance. We also suggest that using error prediction as part of a multi-task learning framework improves performance by allowing the network to compensate for its confidence level. We perform extensive ablation studies to show that each of our contributions increases performance. Furthermore, we show that our approach outperforms the recent state of the art for single-frame 3D human pose estimation by a large margin. Our code and trained models are made publicly available on Github.
△ Less
Submitted 7 August, 2022;
originally announced August 2022.
-
KinePose: A temporally optimized inverse kinematics technique for 6DOF human pose estimation with biomechanical constraints
Authors:
Kevin Gildea,
Clara Mercadal-Baudart,
Richard Blythman,
Aljosa Smolic,
Ciaran Simms
Abstract:
Computer vision/deep learning-based 3D human pose estimation methods aim to localize human joints from images and videos. Pose representation is normally limited to 3D joint positional/translational degrees of freedom (3DOFs), however, a further three rotational DOFs (6DOFs) are required for many potential biomechanical applications. Positional DOFs are insufficient to analytically solve for joint…
▽ More
Computer vision/deep learning-based 3D human pose estimation methods aim to localize human joints from images and videos. Pose representation is normally limited to 3D joint positional/translational degrees of freedom (3DOFs), however, a further three rotational DOFs (6DOFs) are required for many potential biomechanical applications. Positional DOFs are insufficient to analytically solve for joint rotational DOFs in a 3D human skeletal model. Therefore, we propose a temporal inverse kinematics (IK) optimization technique to infer joint orientations throughout a biomechanically informed, and subject-specific kinematic chain. For this, we prescribe link directions from a position-based 3D pose estimate. Sequential least squares quadratic programming is used to solve a minimization problem that involves both frame-based pose terms, and a temporal term. The solution space is constrained using joint DOFs, and ranges of motion (ROMs). We generate 3D pose motion sequences to assess the IK approach both for general accuracy, and accuracy in boundary cases. Our temporal algorithm achieves 6DOF pose estimates with low Mean Per Joint Angular Separation (MPJAS) errors (3.7°/joint overall, & 1.6°/joint for lower limbs). With frame-by-frame IK we obtain low errors in the case of bent elbows and knees, however, motion sequences with phases of extended/straight limbs results in ambiguity in twist angle. With temporal IK, we reduce ambiguity for these poses, resulting in lower average errors.
△ Less
Submitted 6 September, 2022; v1 submitted 26 July, 2022;
originally announced July 2022.
