As machine learning black boxes are increasingly being deployed in domains such as healthcare and criminal justice, there is growing emphasis on building tools and techniques for explaining these black boxes in an interpretable manner. Such explanations are being leveraged by domain experts to diagnose systematic error…
On-device research index
arXiv research
A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
168,742 papers · 148 categories
Trend · papers per month
48 results for “biased input”
New compression methods handle biased input sequences for more accurate posterior summaries.
problem Handling biased input sequences for accurate posterior summaries.
method Stein kernel thinning, low-rank SKT, Stein recombination, Stein Cholesky.
result Achieves accurate posterior summaries with biased input sequences.
New method identifies parameters of wider shallow neural networks with biases.
problem Identifying parameters of wide shallow neural networks with biases from finite samples.
method Two-step pipeline: direction of weights via second order information, signs via algebraic evaluations, biases via gradient descent.
result Constructive methods and theoretical guarantees of finite sample identification for wider shallow networks with biases.
Neural networks with random hidden nodes have gained increasing interest from researchers and practical applications. This is due to their unique features such as very fast training and universal approximation property. In these networks the weights and biases of hidden nodes determining the nonlinear feature mapping a…
A survey of Identification and mitigation of Machine Learning algorithmic biases in Image Analysiscs.LG
Survey on biases in image analysis for industrial safety.
problem Bias in machine learning algorithms affects industrial safety-critical applications.
method Survey and analysis of recent advances in bias detection and mitigation.
result Need for new methods to detect and mitigate biases in image analysis for safety-critical applications.
This work addresses encoding biases in neural networks by tailoring models with unsupervised losses.
problem Improving neural network representations and reducing the generalization gap.
method Inspired by transductive learning, the authors propose tailoring and meta-tailoring to optimize unsupervised losses during prediction time.
result Models trained with tailoring and meta-tailoring perform better on the task objective after adapting to unsupervised losses.
Improves likelihood-free inference by using a new sampling approach to avoid biased data collection.
problem Efficient Bayesian inference without likelihood evaluation for real-world datasets.
method Introduces Neural Proposal (NP) to sample simulation inputs i.i.d. for unbiased posterior inference.
result Demonstrates improved performance, especially for multi-modal posteriors, through experiments.
This paper introduces hierarchical Gaussian process priors for neural networks to capture weight correlations and inductive biases.
problem Capturing weight correlations and inductive biases in neural networks.
method Hierarchical Gaussian process priors with unit embeddings and input-dependent kernels.
result Hierarchical Gaussian process priors provide competitive predictive performance and desirable uncertainty estimates.
Algorithm improves recommendation subset selection in the presence of biases.
problem Maximizing submodular functions for recommendation in the presence of social biases.
method Algorithm for submodular maximization with fairness constraints.
result Algorithm provably outputs subsets with near-optimal utility and proportional representation.
BPN defends against adversarial attacks by generating beneficial perturbations.
problem Adversarial attacks cause deep neural networks to misclassify clean inputs.
method BPN generates beneficial perturbations during training to neutralize future adversarial attacks.
result BPN is robust to adversarial examples and more efficient than classical adversarial training.
Interview study reveals considerations for designing semi-automated bias detection tools.
problem Detecting and mitigating machine learning biases.
method Interview study with 11 machine learning practitioners.
result Four considerations identified for tool design.
Self-attention prefers sparse functions of input sequences, reducing sample complexity.
problem Understanding the inductive biases of self-attention in modeling long-range dependencies.
method Theoretical analysis and synthetic experiments to probe sample complexity of learning sparse functions with Transformers.
result Bounded-norm Transformer networks can represent sparse functions of the input sequence with logarithmic sample complexity.
Recent efforts to improve the interpretability of deep neural networks use saliency to characterize the importance of input features to predictions made by models. Work on interpretability using saliency-based methods on Recurrent Neural Networks (RNNs) has mostly targeted language tasks, and their applicability to tim…
Neural nets learn simple distributions first, then more complex ones.
problem Understanding how neural networks generalize from simple to complex functions.
method Stochastic gradient descent training, synthetic data, CIFAR10, ImageNet pre-training.
result Neural networks initially use lower-order statistics, then higher-order ones.
We prove that the binary classifiers of bit strings generated by random wide deep neural networks with ReLU activation function are biased towards simple functions. The simplicity is captured by the following two properties. For any given input bit string, the average Hamming distance of the closest input bit string wi…
Regularized training of an autoencoder typically results in hidden unit biases that take on large negative values. We show that negative biases are a natural result of using a hidden layer whose responsibility is to both represent the input data and act as a selection mechanism that ensures sparsity of the representati…
Derives equations for deep learning biases and weights, showing data complexity reduction.
problem Understanding interpretability in supervised learning.
method Gradient flow equations and dynamical truncation of training data.
result Data complexity reduction at an exponential rate with training.
Learning algorithms need bias to generalize and perform better than random guessing. We examine the flexibility (expressivity) of biased algorithms. An expressive algorithm can adapt to changing training data, altering its outcome based on changes in its input. We measure expressivity by using an information-theoretic …
Free Random Projection enhances reinforcement learning by naturally incorporating hierarchical structure.
problem Improving reinforcement learning algorithms for better generalization and adaptability.
method Introduces Free Random Projection, a method that uses free probability theory to create random orthogonal matrices encoding hierarchical structure.
result Empirically shows consistent improvement in generalization over standard methods on multi-environment benchmarks.
Eigenvalue Distribution of Large Random Matrices Arising in Deep Neural Networks: Orthogonal Casestat.ML
The paper proves a distribution claim for neural network Jacobians.
problem Distribution of singular values in deep neural networks.
method Free probability and random matrix theory techniques.
result Singular value distribution matches for specific cases.
Algorithm improves binary classification of biased grouped data.
problem Improving binary classification for biased, grouped data.
method Assumes partition-projected class-conditional invariance across groups and derives a semi-supervised algorithm to learn a group-aware classifier.
result Demonstrates improved area under the ROC curve compared to baselines.
We propose an inference method to estimate sparse interactions and biases according to Boltzmann machine learning. The basis of this method is regularization, which is often used in compressed sensing, a technique for reconstructing sparse input signals from undersampled outputs. regularization impedes the …
This paper examines biases in foundation models under long-tailed data and proposes a method to mitigate parameter imbalance.
problem The bias introduced by imbalanced training data in foundation models affects long-tailed downstream tasks.
method The paper examines parameter imbalance and data imbalance, proposing a backdoor adjustment method to mitigate parameter imbalance.
result An average performance increase of about 1.67% on each dataset.
Theoretical analysis of CNNs' inductive biases and their efficiency in approximating functions.
problem Understanding and optimizing the inductive biases in deep CNNs.
method Theoretical analysis combining multichanneling, downsampling, weight sharing, and locality.
result Deep CNNs with depth can approximate any continuous function, and require samples for sparse functions.
Periodic activation functions improve neural network reliability and interpretability.
problem Neural networks reinforce hidden biases, making them unreliable and hard to interpret.
method Introduce periodic activation functions in Bayesian neural networks to establish a connection with stationary Gaussian process priors.
result Periodic activation functions, including sinusoidal, triangular, and ReLU, improve model performance and sensitivity to perturbations.
Transformers are less sensitive to input perturbations compared to other models.
problem Understanding the inductive biases of transformers and distinguishing them from other architectures.
method Identified token-wise sensitivity as a metric to explain transformers' inductive biases across different data modalities.
result Transformers have lower sensitivity than MLPs, CNNs, ConvMixers, and LSTMs, across vision and language tasks.
We analyze the joint probability distribution on the lengths of the vectors of hidden variables in different layers of a fully connected deep network, when the weights and biases are chosen randomly according to Gaussian distributions, and the input is in . We show that, if the activation function sat…
Randomly initialized transformers show extreme token preferences.
problem Structural biases in randomly initialized transformers.
method Dissection of transformer architecture at initialization.
result Initialization-induced biases persist throughout training.
Random deep neural networks are robust to adversarial examples, scaling with input size and dimension.
problem Adversarial examples challenge the reliability of deep learning algorithms.
method Analysis of random deep neural networks with Gaussian process equivalence and experiments on MNIST and CIFAR10.
result The distance of adversarial examples scales as norm of the input.
FANNet: Formal Analysis of Noise Tolerance, Training Bias and Input Sensitivity in Neural Networkscs.LG
FANNet analyzes noise tolerance and training bias in neural networks.
problem Low noise tolerance and input sensitivity in neural networks lead to failures on unseen inputs.
method Formal analysis using model checking under different noise ranges.
result Noise tolerance of for the trained network, sensitive input nodes identified, and biasness confirmed.
Adversarial training, in which a network is trained on both adversarial and clean examples, is one of the most trusted defense methods against adversarial attacks. However, there are three major practical difficulties in implementing and deploying this method - expensive in terms of extra memory and computation costs; …
In many applications, one has side information, e.g., labels that are provided in a semi-supervised manner, about a specific target region of a large data set, and one wants to perform machine learning and data analysis tasks "nearby" that prespecified target region. For example, one might be interested in the clusteri…
Random projections improve GP regression performance, reducing high-dimensional inputs to 1D.
problem Gaussian processes struggle with high-dimensional inputs, leading to overfitting and high computational cost.
method Use additive sums of kernels operating on random projections of inputs.
result Predictive performance converges to full-dimensional kernel performance with increasing projections, even in 1D.
Pointer Graph Networksstat.ML
PGNs dynamically infer and use graph structures to improve model generalization.
problem Static graph structures inferred by machine learning practitioners are often suboptimal for tasks.
method PGNs augment graphs with dynamically inferred pointers for improved model generalization.
result PGNs outperform unrestricted GNNs and Deep Sets on dynamic graph connectivity tasks.
ExpBERT uses natural language explanations to improve text interpretation.
problem Improving text interpretation for relation extraction tasks.
method Fine-tuning BERT on MultiNLI to interpret natural language explanations.
result ExpBERT matches a BERT baseline but requires less labeled data and improves performance.
Applying the maximum entropy principle to neural networks enhances multi-species distribution modelscs.LG
DeepMaxent uses neural networks to improve species distribution models.
problem Sampling biases and lack of absence data in presence-only observations.
method DeepMaxent employs neural networks to learn shared features among species using the maximum entropy principle.
result DeepMaxent outperforms traditional methods in predicting species distributions, especially in unevenly sampled regions.
Deep networks can be biased to learn top eigenfunctions of the kernel outside the training set.
problem Spectral bias of deep networks in the kernel regime.
method Quantitative bounds on difference between finite-width and infinite-width network trajectories.
result Deep networks learn top eigenfunctions of the Neural Tangent Kernel over the entire input space, not just the training set.
This paper proposes using a linear function approximator, rather than a deep neural network (DNN), to bias a Monte Carlo tree search (MCTS) player for general games. This is unlikely to match the potential raw playing strength of DNNs, but has advantages in terms of generality, interpretability and resources (time and …
Machine learning algorithms are optimized to model statistical properties of the training data. If the input data reflects stereotypes and biases of the broader society, then the output of the learning algorithm also captures these stereotypes. In this paper, we initiate the study of gender stereotypes in {\em word emb…
Conditional restricted Boltzmann machines are undirected stochastic neural networks with a layer of input and output units connected bipartitely to a layer of hidden units. These networks define models of conditional probability distributions on the states of the output units given the states of the input units, parame…
This paper studies bounds for the Lipschitz constant of random neural networks.
problem Quantifying the worst-case robustness of neural networks against adversarial perturbations.
method Analyzes upper and lower bounds for the Lipschitz constant of random ReLU neural networks under specific initialization conditions.
result For deep networks, the upper bound is larger than the lower bound by a logarithmic factor in width.
New proof links initial class bias to DNN trainability, challenging traditional understanding.
problem Understanding the initial class bias in DNNs and its impact on trainability.
method Theoretical proof linking initial class bias to mean field theories of DNNs.
result Efficient learning is connected to a network's prejudice towards a specific class, contradicting traditional understanding.
LLMs' explanations are often insufficient and vary with input distribution.
problem Evaluating the sufficiency of LLM explanations without predefined biases.
method Generalizing sufficiency to arbitrary explanations, using LLM's input beliefs, and introducing SCSuff metric.
result Explanation sufficiency can vary with input distribution and is weakly correlated with model size, accuracy, or output entropy.
We consider training machine learning models that are fair in the sense that their performance is invariant under certain sensitive perturbations to the inputs. For example, the performance of a resume screening system should be invariant under changes to the gender and/or ethnicity of the applicant. We formalize this …
Paper explores how knowledge distillation transfers inductive biases between models.
problem Transferring inductive biases between models for tasks with limited data.
method Knowledge distillation applied to models with different inductive biases (LSTMs vs. Transformers, CNNs vs. MLPs).
result Effect of inductive biases is transferred through knowledge distillation, impacting both performance and solution characteristics.
Paper provides unbiased spectral moment estimates from finite data.
problem Challenges in estimating spectral moments from limited data.
method Dynamic programming approach to estimate spectral moments of kernel integral operator.
result Demonstrates consistency with theoretical spectra and practical utility in neural networks.
The standard method of generating random weights and biases in feedforward neural networks with random hidden nodes, selects them both from the uniform distribution over the same fixed interval. In this work, we show the drawbacks of this approach and propose a new method of generating random parameters. This method en…
A new method guides neural networks to focus on specific features of input data.
problem Training neural networks to consider specific features of input data.
method Focus-and-Expand ( ax) method: Gradual manipulation of input features.
result Achieves state-of-the-art results in bias removal and image classification tasks.