Canada

Human values are crucial to human decision-making. \textit{Value pluralism} is the view that multiple correct values may be held in tension with one another (e.g., when considering \textit{lying to a friend to protect their feelings}, how does one balance \textit{honesty} with \textit{friendship}?). As statistical learners, AI systems fit to averages by default, washing out these potentially irreducible value conflicts. To improve AI systems to better reflect value pluralism, the first-order challenge is to explore the extent to which AI systems can model pluralistic human \textit{values}, \textit{rights}, and \textit{duties} as well as their interaction.

We introduce ValuePrism, a large-scale dataset of 218k values, rights, and duties connected to 31k human-written situations. ValuePrism&#39;s contextualized values are generated by GPT-4 and deemed high-quality by human annotators 91% of the time. We conduct a large-scale study with annotators across diverse social and demographic backgrounds to try to understand whose values are represented.

With ValuePrism, we build Value Kaleidoscope (or Kaleido), an open, light-weight, and structured language-based multi-task model that generates, explains, and assesses the relevance and valence (i.e., support or oppose) of human values, rights, and duties within a specific context. Humans prefer the sets of values output by our system over the teacher GPT-4, finding them more accurate and with broader coverage. In addition, we demonstrate that Kaleido can help explain variability in human decision-making by outputting contrasting values. Finally, we show that Kaleido&#39;s representation transfer to other philosophical frameworks and datasets, confirming the benefit of an explicit, modular, and interpretable approach to value pluralism. We hope that our work will serve as a step to making more explicit the implicit values behind human decision-making and to steering AI systems to make decisions that are more in accordance with them.

AAAI 2024

Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties

pluralism

peai: morality and value-based ai

snlp: applications

human values

morality

alignment

philosophy

natural language processing

Human values are crucial to human decision-making. \textit{Value pluralism} is the view that multiple correct values may be held in tension with one another (e.g., when considering \textit{lying to a friend to protect their feelings}, how does one balance \textit{honesty} with \textit{friendship}?). As statistical learners, AI systems fit to averages by default, washing out these potentially irreducible value conflicts. To improve AI systems to better reflect value pluralism, the first-order challenge is to explore the extent to which AI systems can model pluralistic human \textit{values}, \textit{rights}, and \textit{duties} as well as their interaction.

We introduce ValuePrism, a large-scale dataset of 218k values, rights, and duties connected to 31k human-written situations. ValuePrism's contextualized values are generated by GPT-4 and deemed high-quality by human annotators 91% of the time. We conduct a large-scale study with annotators across diverse social and demographic backgrounds to try to understand whose values are represented.

With ValuePrism, we build Value Kaleidoscope (or Kaleido), an open, light-weight, and structured language-based multi-task model that generates, explains, and assesses the relevance and valence (i.e., support or oppose) of human values, rights, and duties within a specific context. Humans prefer the sets of values output by our system over the teacher GPT-4, finding them more accurate and with broader coverage. In addition, we demonstrate that Kaleido can help explain variability in human decision-making by outputting contrasting values. Finally, we show that Kaleido's representation transfer to other philosophical frameworks and datasets, confirming the benefit of an explicit, modular, and interpretable approach to value pluralism. We hope that our work will serve as a step to making more explicit the implicit values behind human decision-making and to steering AI systems to make decisions that are more in accordance with them.

technical paper

We are pleased to announce the Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI-24), which will be held in Vancouver, British Columbia at the Vancouver Convention Centre – West Building from 20-27 February, 2024.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-24 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

We expect for AAAI-24 to be an in-person conference – one author of all accepted papers will be expected to present work in person unless there are exceptional circumstances that prevent this.<br><br><br><br>

In order to access the AAAI-24 event page you need to register [here](https://aaai.org/aaai-conference/registration/)

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. 

Many inductive logic programming approaches struggle to learn programs from noisy data. To overcome this limitation, we introduce an approach that learns minimal description length programs from noisy data, including recursive programs. Our experiments on several domains, including drug design, game playing, and program synthesis, show that our approach can outperform existing approaches in terms of predictive accuracies and scale to moderate amounts of noise.

Learning MDL Logic Programs from Noisy Data

We study the problem of agent selection in causal strategic learning under multiple decision makers and address two key challenges that come with it. 
Firstly, while much of prior work focuses on studying a fixed pool of agents that remains static regardless of their evaluations, we consider the impact of selection procedure by which agents are not only evaluated, but also selected.
When each decision maker unilaterally selects agents by maximising their own utility, we show that the optimal selection rule is a trade-off between selecting the best agents and providing incentives to maximise the agents' improvement. 
Furthermore, this optimal selection rule relies on incorrect predictions of agents' outcomes. 
Hence, we study the conditions under which a decision maker's optimal selection rule will not lead to deterioration of agents' outcome nor cause unjust reduction in agents' selection chance. 
To that end, we provide an analytical form of the optimal selection rule and a mechanism to retrieve the causal parameters from observational data, under certain assumptions on agents' behaviour. 
Secondly, when there are multiple decision makers, the interference between selection rules introduces another source of biases in estimating the underlying causal parameters. 
To address this problem, we provide a cooperative protocol which all decision makers must collectively adopt to recover the true causal parameters. 
Lastly, we complement our theoretical results with simulation studies.
Our results highlight not only the importance of causal modeling as a strategy to mitigate the effect of gaming, as suggested by previous work, but also the need of a benevolent regulator to enable it.

Causal Strategic Learning with Competitive Selection

Domain-independent dynamic programming (DIDP), a model-based paradigm based on dynamic programming, has shown promising performance on multiple combinatorial optimization problems compared with mixed integer programming (MIP) and constraint programming (CP). The current DIDP solvers are based on heuristic search, and the state-of-the-art solver, complete anytime beam search (CABS), uses beam search. However, the current DIDP solvers cannot utilize multiple threads, unlike state-of-the-art MIP and CP solvers. In this paper, we propose three parallel beam search algorithms and develop multi-thread implementations of CABS. With 32 threads, our multi-thread DIDP solvers achieve 9 to 39 times speedup on average and significant performance improvement over the sequential solver, finding the new best solutions for two instances of the traveling salesperson problem with time windows. In addition, our solvers outperform multi-thread MIP and CP solvers in four of the six combinatorial optimization problems evaluated.

Parallel Beam Search Algorithms for Domain-Independent Dynamic Programming | VIDEO

As machine learning models, specifically neural networks, are becoming increasingly popular, there are concerns regarding their trustworthiness, specially in safety-critical applications, e.g. actions of an autonomous vehicle must be safe. There are approaches that can train neural networks where such domain requirements are enforced as constraints, but they either cannot guarantee that the constraint will be satisfied by all possible predictions (even on unseen data) or they are limited in the type of constraints that can be enforced. In this paper, we present an approach to train neural networks which can enforce a wide variety of constraints and guarantee that the constraint is satisfied by all possible predictions. The approach builds on earlier work where learning linear models is formulated as a constraint satisfaction problem (CSP). To make this idea applicable to neural networks, two crucial new elements are added: constraint propagation over the network layers, and weight updates based on a mix of gradient descent and CSP solving. Evaluation on various machine learning tasks demonstrates that our approach is flexible enough to enforce a wide variety of domain constraints and is able to guarantee them in neural networks.

DeepSaDe: Learning Neural Networks That Guarantee Domain Constraint Satisfaction

Boolean Networks (BNs) are widely used as a modeling formalism in several domains, notably systems biology and computer science. A fundamental problem in BN analysis is the enumeration of trap spaces, which are hypercubes in the state space that cannot be escaped once entered. Several methods have been proposed for enumerating trap spaces, however they often suffer from scalability and efficiency issues, particularly for large and complex models. To our knowledge, the most efficient and recent methods for the trap space enumeration all rely on Answer Set Programming (ASP), which has been widely applied to the analysis of BNs. Motivated by these considerations, our work proposes a new method for enumerating trap spaces in BNs using ASP. We evaluate the method on a mix of 250+ real-world and 400+ randomly generated BNs, showing that it enables analysis of models beyond the capabilities of existing tools (namely pyboolnet, mpbn, trappist, and trapmvn).

Scalable Enumeration of Trap Spaces in Boolean Networks via Answer Set Programming

Virtual Reality (VR) has emerged due to advancements in hardware and computer graphics. During the pandemic, conferences and exhibitions leveraging VR have gained attention. However, large-scale VR conferences, face a significant problem not yet studied in the
literature -- displaying too many irrelevant users on the screen which may negatively impact the user experience. To address this issue, we formulate a new research problem, {\it \underline{S}ocial-Aware \underline{V}R Conference \underline{G}roup \underline{D}isplay Configuration (SVGD)}. Accordingly, we design the {\it \underline{S}ocial Utility-Aware \underline{V}R \underline{C}onference Group Formation (SVC)} algorithm, which is a $2$-approximation algorithm to SVGD. We perform a user study on SVGD, and the results show that the configuration identified by SVC receives the highest satisfaction. Experiments on real metaverse datasets show that the proposed SVC outperforms 11 baselines by 75\% in terms of solution quality.

Social-Aware Group Display Configuration in VR Conference

Graphs are typical non-Euclidean data of complex structures. Recently, Riemannian graph representation learning emerges as an exciting alternative to the traditional Euclidean ones. However, Riemannian graph methods are still in an early stage: most of them present a single curvature (radius) regardless of structural complexity, suffer from numerical instability due to the exponential/logarithmic map, and lack the ability to capture motif regularity. In light of the issues above, we propose the problem of Motif-aware Riemannian Graph Representation Learning, seeking a numerically stable encoder to capture motif regularity in a diverse-curvature manifold without labels. To this end, we present a novel Motif-aware Riemannian model with Generative-Contrastive learning (MotifRGC), which conducts a minmax game in Riemannian manifold in a self-supervised manner. First, we propose a new type of Riemannian GCN (D-GCN), in which we construct a diverse-curvature manifold by a product layer with a diversified factor, and replace the exponential/logarithmic map by a stable kernel layer proved to be isometric invariant. Second, we introduce a motif-aware Riemannian generative-contrastive learning to capture motif regularity in the manifold and learn motif-aware node representation without external labels. Extensive empirical results show the superiority of MofitRGC against prior Riemannian methods.

Motif-Aware Riemannian Graph Neural Network with Generative-Contrastive Learning

Decision Trees (DTs) are commonly used for many machine learning tasks due to their high degree of interpretability. However, learning a DT from data is a difficult optimization problem, as it is non-convex and non-differentiable. Therefore, common approaches learn DTs using a greedy growth algorithm that minimizes the impurity locally at each internal node. Unfortunately, this greedy procedure can lead to inaccurate trees. In this paper, we present a novel approach for learning hard, axis-aligned DTs with gradient descent. The proposed method uses backpropagation with a straight-through operator on a dense DT representation, to jointly optimize all tree parameters. Our approach outperforms existing methods on binary classification benchmarks and achieves competitive results for multi-class tasks. The implementation is available under: https://github.com/s-marton/GradTree

GradTree: Learning Axis-Aligned Decision Trees with Gradient Descent

A major limitation for the broader scope of problems solvable by transformers is the quadratic scaling of computational complexity with input size. In this study, we investigate the recurrent memory augmentation of pre-trained transformer models to extend input context length while linearly scaling compute. Our approach demonstrates the capability to store information in memory for sequences of up to an unprecedented two million tokens while maintaining high retrieval accuracy. Experiments with language modeling tasks show perplexity improvement as the number of processed input segments increases. These results underscore the effectiveness of our method, which has significant potential to enhance long-term dependency handling in natural language understanding and generation tasks, as well as enable large-scale context processing for memory-intensive applications.

Beyond Attention: Breaking the Limits of Transformer Context Length with Recurrent Memory | VIDEO

Anomaly detection (AD), also termed outlier detection, is a pivotal technique in machine learning (ML) and artificial intelligence (AI). In today’s interconnected landscape, AD’s importance is evident across diverse sectors, from national security to healthcare and security. For instance, NASA uses AD algorithms for spacecraft failure detection, the U.S. Navy applies it for enhanced space system autonomy, the University of Pittsburgh Medical Center uses AD for early patient prioritization, and companies like Google and Amazon employ AD for service disruption identification. Fundamentally, AD is invaluable in identifying rare yet significant data points in large datasets. My research focuses on using algorithmic and system methods to automate and accelerate AD. The goal is to develop automated, scalable, and reproducible detection algorithms and systems to support high-stakes research and applications. To sum up, my research has produced successful results: (1) Automated algorithms; (2) Scalable AD systems and (3) Reproducible AD benchmarks, which will be covered in the proposed presentation. Please see the attached cover letter and presentation schedule for details.

Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties

Downloads

Next from AAAI 2024

Learning MDL Logic Programs from Noisy Data

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAAI 2024

Learning MDL Logic Programs from Noisy Data

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads