RESTA defends LLMs against jailbreaking attacks by adding random noise to embeddings.
problem Vulnerability of LLMs to jailbreaking attacks that generate harmful outputs.
method Adds random noise to embedding vectors and aggregates during token generation.
result RESTA achieves superior robustness versus utility tradeoffs compared to baseline defenses.
Paper introduces Latent-CLIP for efficient text-image comparison in latent space.
problem Efficiently compare text and images in latent space without costly decoding.
method Trains CLIP model in latent space, uses Latent-CLIP rewards for noise optimization, and guides generation away from harmful content.
result Latent-CLIP matches CLIP performance on text-image classification and harmful content detection.
Automatically jailbreaks LLMs with black-box access.
problem Generating harmful content from black-box LLMs.
method Automated method using an attacker LLM to refine prompts.
result Generates jailbreaks for over 80% of prompts.
The large volume of text in electronic healthcare records often remains underused due to a lack of methodologies to extract interpretable content. Here we present an unsupervised framework for the analysis of free text that combines text-embedding with paragraph vectors and graph-theoretical multiscale community detect…
Proposes a new method to measure and avoid harm in machine learning decisions.
problem Measuring and avoiding harm in machine learning algorithms.
method Formal definition of harm and benefit using causal models, counterfactual objective functions.
result Demonstrates that standard machine learning methods can lead to harmful policies under distributional shifts.
Contrastive learning harms minority group representations, affecting downstream tasks.
problem Representation harm in contrastive learning, especially affecting minority groups.
method Causal mediation analysis and stochastic block model explanation.
result Representation harm in contrastive learning is partly responsible for allocation harm in downstream tasks.
While harms of allocation have been increasingly studied as part of the subfield of algorithmic fairness, harms of representation have received considerably less attention. In this paper, we formalize two notions of stereotyping and show how they manifest in later allocative harms within the machine learning pipeline. …
Aims to create safe reinforcement learning policies by considering individual harm.
problem Optimal policies for a population may harm certain individuals.
method Formalizes individual harm, proposes a two-stage procedure, and establishes finite-sample properties.
result Learned policies maximize expected return while minimizing harm.
Researchers classify harmful structures on unimodular Lie groups.
problem Characterizing harmful structures on Lie groups.
method Analyzing left-invariant structures and using Clifford multiplication.
result Classification of harmful structures on unimodular Lie groups of dimension ≤ 4.
This paper formalizes AI safety using hypothesis testing in GenAI.
problem Ensuring safety of generative AI tools that create realistic content.
method Formalization of computational safety through hypothesis testing and signal processing.
result Demonstrates how AI safety can be assessed quantitatively using mathematical frameworks.
Prediction models can harm patients even when accurate, leading to self-fulfilling prophecies.
problem Prediction models can lead to harmful decisions that worsen patient outcomes.
method Formal characterization of harmful prediction models and analysis of their impact.
result Well-calibrated models are ineffective for decision-making as they do not change the data distribution.
New method stops experiments early for harm in diverse groups.
problem Early stopping of experiments for harmful treatment effects in diverse populations.
method Causal machine learning approach (CLASH) for early stopping.
result CLASH effectively stops experiments early for harmful treatment effects in diverse groups.
The paper proves neural networks are almost always surjective, impacting model safety.
problem Ensuring neural networks can generate any output, including harmful content.
method Analyzing fundamental neural architectures and generative models.
result Many neural architectures are almost always surjective, allowing for arbitrary outputs.
The paper investigates ethical issues in large image datasets, focusing on pornographic content.
problem Ethical issues in large-scale computer vision datasets, particularly concerning pornographic content.
method Cross-sectional model-based quantitative census covering various factors in the ImageNet-ILSVRC-2012 dataset.
result The dataset contains verifiably pornographic images, including non-consensual and voyeuristic content.
Theorem. Let M be a compact, connected, oriented smooth Riemannian n-manifold with non-empty boundary. Then the cohomology of the complex (Harm*(M),d) of harmonic forms on M is given by the direct sum H^p(Harm*(M),d) = H^p(M;R) + H^(p-1)(M;R) for p=0,1,...,n. When M is a closed manifold, a form is harmonic if and only …
The viral spread of fake news has caused great social harm, making fake news detection an urgent task. Current fake news detection methods rely heavily on text information by learning the extracted news content or writing style of internal knowledge. However, deliberate rumors can mask writing style, bypassing language…
New method estimates harmful instances in GANs for better model performance.
problem Estimating harmful instances in GANs for improved model performance.
method Uses Jacobian of generator's loss to trace harmfulness of instances.
result Correctly inferred changes in GAN evaluation metrics and improved model performance.
New attacks reduce bad queries in black-box classifiers, improving effectiveness.
problem Evasion attacks often issue many bad queries, which are costly.
method Design new attacks that minimize bad queries while increasing non-bad queries.
result Reduces the number of bad queries by 1.5-7.3 times, improving attack effectiveness.
Study shows harmful overfitting in Sobolev spaces even as training data grows.
problem Harmful overfitting in Sobolev spaces under noisy conditions.
method Geometric argument using Sobolev inequalities.
result Approximately norm-minimizing interpolators exhibit harmful overfitting.
New framework shows algorithmic recourse can be harmful.
problem Risks of providing algorithmic recourse in binary classification.
method Abstract learning-theoretic framework comparing risks with and without recourse.
result Providing recourse can be harmful, increasing class uncertainty and mistakes.
As machine learning (ML) increasingly affects people and society, awareness of its potential unwanted consequences has also grown. To anticipate, prevent, and mitigate undesirable downstream consequences, it is critical that we understand when and how harm might be introduced throughout the ML life cycle. In this paper…
Common fairness definitions in machine learning focus on balancing notions of disparity and utility. In this work, we study fairness in the context of risk disparity among sub-populations. We are interested in learning models that minimize performance discrepancies across sensitive groups without causing unnecessary ha…
Improves GAN performance by identifying and removing harmful training instances.
problem Challenges in applying previous instance removal methods to GANs.
method Proposes influence estimation methods using Jacobian of generator's gradient and instance evaluation scheme based on GAN evaluation metrics.
result Removing harmful instances significantly improves GAN performance.
Bayesian approach quantifies uncertainty in LLM evaluations.
problem Statistical uncertainty in evaluating LLM behavior.
method Bayesian evaluation of LLM behavior using probabilistic text generation strategies.
result Bayesian approach provides useful uncertainty quantification about LLM behavior.
Study shows online learning algorithms incentivize low-quality content, proposing new algorithms to improve quality.
problem Online learning algorithms in content recommender systems incentivize producers to create low-quality content.
method Analyzed the game between producers and content quality, designed new learning algorithms to incentivize high effort and quality.
result New algorithms incentivize producers to invest high effort and achieve high user welfare, improving content quality.
Proposes a model to optimize feedback for content creators on social media.
problem Nurturing content creation on social media platforms.
method Modeling feedback distribution to incentivize creators and optimize newsfeed experience.
result Optimized newsfeed for content creators without compromising user experience.
Proposes a VAE variant for ordinal content factors.
problem Isolating ordinal-valued content factors in deep latent variable models.
method Introduces a partially ordered set (poset) structure and a conditional Gaussian spacing prior model.
result Significant improvements in content-style separation over previous non-ordinal approaches.
The study finds that memorization is necessary or harmful depending on the prior distribution and noise level.
problem The impact of memorization on generalization in overparameterized models.
method An overparameterized linear model with general priors in a Bayesian setup.
result Explicit conditions for optimal generalization based on the prior distribution and noise level.
Sensory data are often comprised of independent content and transformation factors. For example, face images may have shapes as content and poses as transformation. To infer separately these factors from given data, various ``disentangling'' models have been proposed. However, many of these are supervised or semi-super…
New framework assesses LLM security risks in BFSI.
problem Lack of domain-specific security evaluation for LLMs in BFSI.
method Risk-aware evaluation framework combining taxonomy, automated red-teaming, and ensemble judging.
result Higher decoding stochasticity and adaptive interaction lead to more severe disclosures.
In the industry of video content providers such as VOD and IPTV, predicting the popularity of video contents in advance is critical not only from a marketing perspective but also from a network optimization perspective. By predicting whether the content will be successful or not in advance, the content file, which is l…
Many businesses are using recommender systems for marketing outreach. Recommendation algorithms can be either based on content or driven by collaborative filtering. We study different ways to incorporate content information directly into the matrix factorization approach of collaborative filtering. These content-booste…
Spotify improves content mix using contextual bandits.
problem Skewed historical data and varying user preferences across contexts.
method Contextual bandits to dynamically learn optimal content type distribution.
result Improved precision and user engagement with under-represented content types.
A new model considers fatigue in online content recommendation systems.
problem Fatigue in users due to overexposure and boredom from similar recommendations.
method Proposed a fatigue-aware Dependent Click Model (DCM) and two learning algorithms.
result Developed algorithms with regret bounds for learning content relevance and fatigue effects.
Study characterizes harmful low-fidelity data sources for surrogate models.
problem Identifying which low-fidelity data sources to use in constructing surrogate models.
method Employed benchmark filtering techniques to assess harmful sources using limited data.
result Provided guidelines for using low-fidelity sources in an industrial setting.
New PCGML approach generates novel game content across multiple platformer domains.
problem Generating novel game content in new domains.
method Using a new affordance and path vocabulary, variational autoencoders trained on data from six platformer games produce new content with varying proportions of different domains.
result Captures latent level space spanning multiple domains and generates new content with varying proportions of different domains.
Islamophobic hate speech on social media inflicts considerable harm on both targeted individuals and wider society, and also risks reputational damage for the host platforms. Accordingly, there is a pressing need for robust tools to detect and classify Islamophobic hate speech at scale. Previous research has largely ap…
We consider visual domains in which a class label specifies the content of an image, and class-irrelevant properties that differentiate instances constitute the style. We present a domain-independent method that permits the open-ended recombination of style of one image with the content of another. Open ended simply me…
New study shows personalized content recommendations can lead to polarization of user preferences.
problem Personalized content recommendations can alter user preferences, leading to polarization.
method Used a model of preference dynamics to explore how personalized content affects user preferences.
result Standard reward maximization algorithms achieve only constant regret in personalized recommendation environments.
Neural Networks are being integrated into safety critical systems, e.g., perception systems for autonomous vehicles, which require trained networks to perform safely in novel scenarios. It is challenging to verify neural networks because their decisions are not explainable, they cannot be exhaustively tested, and finit…
Text style transfer aims to modify the style of a sentence while keeping its content unchanged. Recent style transfer systems often fail to faithfully preserve the content after changing the style. This paper proposes a structured content preserving model that leverages linguistic information in the structured fine-gra…
Study on heat content for submanifolds in sub-Riemannian geometry.
problem Understanding heat content for submanifolds in sub-Riemannian geometry.
method Existence of smooth tubular neighborhood, definition of relative heat content, approximation via smooth neighborhoods, asymptotic expansion analysis.
result Approximation of relative heat content fails to recover the exact expansion.
With rapid development of the Internet, web contents become huge. Most of the websites are publicly available, and anyone can access the contents from anywhere such as workplace, home and even schools. Nevertheless, not all the web contents are appropriate for all users, especially children. An example of these content…
Introduces MPR to measure and optimize representation across intersectional groups in retrieval.
problem Harmful stereotypes, cultural erasure, and social disparities in image search and retrieval.
method Develops MPR metric, practical estimation methods, theoretical guarantees, and optimization algorithms.
result Optimizing MPR yields more proportional representation across multiple intersectional groups, often with minimal retrieval accuracy compromise.
Enhances content moderation with culturally-aware models.
problem Global content moderation policies miss local cultural nuances.
method Fine-tuning encoder-decoder models on media-diet data.
result Improved accuracy in local violation detection and cultural alignment.
Much of the data being created on the web contains interactions between users and items. Stochastic blockmodels, and other methods for community detection and clustering of bipartite graphs, can infer latent user communities and latent item clusters from this interaction data. These methods, however, typically ignore t…
Transforms web content for better visibility in AI-driven search engines.
problem Disruption of traditional SEO by generative AI search engines.
method Fine-tunes a BART-base transformer on synthetically generated training data.
result Significant improvements in ROUGE-L and BLEU scores, and substantial visibility gains in generative search responses.
Improves content allocation in educational platforms with sparse data.
problem Imbalanced content allocation and delayed convergence in adaptive strategies.
method Introduces WAPTS, an algorithm that refines Thompson Sampling for data-sparse environments.
result Demonstrates earlier and more reliable identification of promising treatments.