A former senior employee at OpenAI has said the company behind ChatGPT is prioritising "shiny products" over safety, revealing that he quit after a disagreement over key aims reached "breaking point". Jan Leike was a key safety researcher at OpenAI as its co-head of superalignment, ensuring that powerful artificial intelligence systems adhere to human values and aims. His intervention comes before a global artificial intelligence summit in Seoul next week, where politicians, experts and tech executives will discuss oversight of the technology. Leike resigned days after the San Francisco-based company launched its latest AI model, GPT-4o. His departure means two senior safety figures at OpenAI have left this week following the resignation of Ilya Sutskever, OpenAI's co-founder and fellow co-head of superalignment.

In this paper, we propose a structure-guided Gauss-Newton (SgGN) method for solving least squares problems using a shallow ReLU neural network. The method effectively takes advantage of both the least squares structure and the neural network structure of the objective function. By categorizing the weights and biases of the hidden and output layers of the network as nonlinear and linear parameters, respectively, the method iterates back and forth between the nonlinear and linear parameters. The nonlinear parameters are updated by a damped Gauss-Newton method and the linear ones are updated by a linear solver. Moreover, at the Gauss-Newton step, a special form of the Gauss-Newton matrix is derived for the shallow ReLU neural network and is used for efficient iterations. It is shown that the corresponding mass and Gauss-Newton matrices in the respective linear and nonlinear steps are symmetric and positive definite under reasonable assumptions. Thus, the SgGN method naturally produces an effective search direction without the need of additional techniques like shifting in the Levenberg-Marquardt method to achieve invertibility of the Gauss-Newton matrix. The convergence and accuracy of the method are demonstrated numerically for several challenging function approximation problems, especially those with discontinuities or sharp transition layers that pose significant challenges for commonly used training algorithms in machine learning.

Ensuring a neural network is not relying on protected attributes (e.g., race, sex, age) for predictions is crucial in advancing fair and trustworthy AI. While several promising methods for removing attribute bias in neural networks have been proposed, their limitations remain under-explored. In this work, we mathematically and empirically reveal an important limitation of attribute bias removal methods in presence of strong bias. Specifically, we derive a general non-vacuous information-theoretical upper bound on the performance of any attribute bias removal method in terms of the bias strength. We provide extensive experiments on synthetic, image, and census datasets to verify the theoretical bound and its consequences in practice. Our findings show that existing attribute bias removal methods are effective only when the inherent bias in the dataset is relatively weak, thus cautioning against the use of these methods in smaller datasets where strong attribute bias can occur, and advocating the need for methods that can overcome this limitation.

Ensuring a neural network is not relying on protected attributes (e.g., race, sex, age) for prediction is crucial in advancing fair and trustworthy AI. While several promising methods for removing attribute bias in neural networks have been proposed, their limitations remain under-explored. To that end, in this work, we mathematically and empirically reveal the limitation of existing attribute bias removal methods in presence of strong bias and propose a new method that can mitigate this limitation. Specifically, we first derive a general non-vacuous information-theoretical upper bound on the performance of any attribute bias removal method in terms of the bias strength, revealing that they are effective only when the inherent bias in the dataset is relatively weak. Next, we derive a necessary condition for the existence of any method that can remove attribute bias regardless of the bias strength. Inspired by this condition, we then propose a new method using an adversarial objective that directly filters out protected attributes in the input space while maximally preserving all other attributes, without requiring any specific target label. The proposed method achieves state-of-the-art performance in both strong and moderate bias settings. We provide extensive experiments on synthetic, image, and census datasets, to verify the derived theoretical bound and its consequences in practice, and evaluate the effectiveness of the proposed method in removing strong attribute bias.

The World Health Organization has classified compulsive game playing as a mental health condition. Jones, 30, is charged with four counts of reckless endangerment involving a deadly weapon, all felonies, in the incident a little after 1:30 a.m. One of the bullets struck a house across the street, and a family of three was at home, arrest warrants state. Authorities later found the bullet lodged behind a window shutter. No one inside the house was injured.

This week, ServiceNow released the results of a new report, revealing that a majority of organisations have introduced advanced automation in their workplace – with the full effects being felt next year. A third of companies said that by 2018 they will need greater automation to handle the volume of tasks being generated. By 2020, eight out of 10 UK companies will hit that breaking point which is being contributed to by the overload created by mobile devices and the Internet of Things (IoT). Intelligent automation could address the issue of the breaking point and increase productivity. This includes artificial intelligence or machine learning to streamline decision making to improve the speed and accuracy of business processes.

We investigate the problem of revenue optimal mechanism design [Myerson, 1981] under the context of the partial rationality model, where buyers randomize between two modes: rational and irrational. When a buyer is irrational (can be thought of as lazy), he acts according to certain fixed strategies, such as bidding his true valuation. The seller cannot observe the buyer’s valuation, or his rationality mode, but treat them as random variables from known distributions. The seller’s goal is to design a single-shot auction that maximizes her expected revenue. A minor generalization as it may seem, our findings are in sharp contrast to Myerson’s theory on the standard rational bidder case. In particular, we show that, even for the simplest setting with one buyer, direct value revelation loses generality. However, we do show that, in terms of revenue, the optimal value-revelation and type-revelation mechanisms are equivalent. In addition, the posted-price mechanism is no longer optimal. In fact, the more complicated the mechanism, the higher the revenue. For the case where there are multiple bidders with IID uniform valuations, we show that when the irrational buyers are truthful, first price auction yields more revenue than second price auction.

Variable selection in linear models plays a pivotal role in modern statistics. Hard-thresholding methods such as $l_0$ regularization are theoretically ideal but computationally infeasible. In this paper, we propose a new approach, called the LAGS, short for "least absulute gradient selector", to this challenging yet interesting problem by mimicking the discrete selection process of $l_0$ regularization. To estimate $\beta$ under the influence of noise, we consider, nevertheless, the following convex program [\hat{\beta} = \textrm{arg min}\frac{1}{n}\|X^{T}(y - X\beta)\|_1 + \lambda_n\sum_{i = 1}^pw_i(y;X;n)|\beta_i|] $\lambda_n > 0$ controls the sparsity and $w_i > 0$ dependent on $y, X$ and $n$ is the weights on different $\beta_i$; $n$ is the sample size. Surprisingly, we shall show in the paper, both geometrically and analytically, that LAGS enjoys two attractive properties: (1) LAGS demonstrates discrete selection behavior and hard thresholding property as $l_0$ regularization by strategically chosen $w_i$, we call this property "pseudo-hard thresholding"; (2) Asymptotically, LAGS is consistent and capable of discovering the true model; nonasymptotically, LAGS is capable of identifying the sparsity in the model and the prediction error of the coefficients is bounded at the noise level up to a logarithmic factor---$\log p$, where $p$ is the number of predictors. Computationally, LAGS can be solved efficiently by convex program routines for its convexity or by simplex algorithm after recasting it into a linear program. The numeric simulation shows that LAGS is superior compared to soft-thresholding methods in terms of mean squared error and parsimony of the model.