We use cookies to analyze the browsing and usage of our website and to personalize your experience. You can disable these technologies at any time, but this may limit certain functionalities of the site. Read our Privacy Policy for more information.
Setting cookies
You can enable and disable the types of cookies you wish to accept. However certain choices you make could affect the services offered on our sites (e.g. suggestions, personalised ads, etc.).
Essential cookies
These cookies are necessary for the operation of the site and cannot be deactivated. (Still active)
Analytics cookies
Do you accept the use of cookies to measure the audience of our sites?
Multimedia Player
Do you accept the use of cookies to display and allow you to watch the video content hosted by our partners (YouTube, etc.)?
Governance frameworks should address the prospect of AI systems that cannot be safely tested Technical experts and policy-makers have increa… (see more)singly emphasized the need to address extinction risk from artificial intelligence (AI) systems that might circumvent safeguards and thwart attempts to control them (1). Reinforcement learning (RL) agents that plan over a long time horizon far more effectively than humans present particular risks. Giving an advanced AI system the objective to maximize its reward and, at some point, withholding reward from it, strongly incentivizes the AI system to take humans out of the loop, if it has the opportunity. The incentive to deceive humans and thwart human control arises not only for RL agents but for long-term planning agents (LTPAs) more generally. Because empirical testing of sufficiently capable LTPAs is unlikely to uncover these dangerous tendencies, our core regulatory proposal is simple: Developers should not be permitted to build sufficiently capable LTPAs, and the resources required to build them should be subject to stringent controls.
Governance frameworks should address the prospect of AI systems that cannot be safely tested Technical experts and policy-makers have increa… (see more)singly emphasized the need to address extinction risk from artificial intelligence (AI) systems that might circumvent safeguards and thwart attempts to control them (1). Reinforcement learning (RL) agents that plan over a long time horizon far more effectively than humans present particular risks. Giving an advanced AI system the objective to maximize its reward and, at some point, withholding reward from it, strongly incentivizes the AI system to take humans out of the loop, if it has the opportunity. The incentive to deceive humans and thwart human control arises not only for RL agents but for long-term planning agents (LTPAs) more generally. Because empirical testing of sufficiently capable LTPAs is unlikely to uncover these dangerous tendencies, our core regulatory proposal is simple: Developers should not be permitted to build sufficiently capable LTPAs, and the resources required to build them should be subject to stringent controls.
Abstract Sentences containing multiple semantic operators with overlapping scope often create ambiguities in interpretation, known as scope … (see more)ambiguities. These ambiguities offer rich insights into the interaction between semantic structure and world knowledge in language processing. Despite this, there has been little research into how modern large language models treat them. In this paper, we investigate how different versions of certain autoregressive language models—GPT-2, GPT-3/3.5, Llama 2, and GPT-4—treat scope ambiguous sentences, and compare this with human judgments. We introduce novel datasets that contain a joint total of almost 1,000 unique scope-ambiguous sentences, containing interactions between a range of semantic operators, and annotated for human judgments. Using these datasets, we find evidence that several models (i) are sensitive to the meaning ambiguity in these sentences, in a way that patterns well with human judgments, and (ii) can successfully identify human-preferred readings at a high level of accuracy (over 90% in some cases).1
Abstract Sentences containing multiple semantic operators with overlapping scope often create ambiguities in interpretation, known as scope … (see more)ambiguities. These ambiguities offer rich insights into the interaction between semantic structure and world knowledge in language processing. Despite this, there has been little research into how modern large language models treat them. In this paper, we investigate how different versions of certain autoregressive language models—GPT-2, GPT-3/3.5, Llama 2, and GPT-4—treat scope ambiguous sentences, and compare this with human judgments. We introduce novel datasets that contain a joint total of almost 1,000 unique scope-ambiguous sentences, containing interactions between a range of semantic operators, and annotated for human judgments. Using these datasets, we find evidence that several models (i) are sensitive to the meaning ambiguity in these sentences, in a way that patterns well with human judgments, and (ii) can successfully identify human-preferred readings at a high level of accuracy (over 90% in some cases).1
The jar test is the current standard method for predicting the performance of a conventional drinking water treatment (DWT) process and opti… (see more)mizing the coagulant dose. This test is time-consuming and requires human intervention, meaning it is infeasible for making continuous process predictions. As a potential alternative, we developed a machine learning (ML) model from historical DWT plant data that can operate continuously using real-time sensor data without human intervention for predicting clarified water turbidity 15 min in advance. We evaluated three types of models: multilayer perceptron (MLP), the long short-term memory (LSTM) recurrent neural network (RNN), and the gated recurrent unit (GRU) RNN. We also employed two training methodologies: the commonly used holdout method and the theoretically correct blocked cross-validation (BCV) method. We found that the RNN with GRU was the best model type overall and achieved a mean absolute error on an independent production set of as low as 0.044 NTU. We further found that models trained using BCV typically achieve errors equal to or lower than their counterparts trained using holdout. These results suggest that RNNs trained using BCV are superior for the development of ML models for DWT processes compared to those reported in earlier literature.
Suicide is a complex, multidimensional event, and a significant challenge for prevention globally. Artificial intelligence (AI) and machine … (see more)learning (ML) have emerged to harness large-scale datasets to enhance risk detection. In order to trust and act upon the predictions made with ML, more intuitive user interfaces must be validated. Thus, Interpretable AI is one of the crucial directions which could allow policy and decision makers to make reasonable and data-driven decisions that can ultimately lead to better mental health services planning and suicide prevention. This research aimed to develop sex-specific ML models for predicting the population risk of suicide and to interpret the models. Data were from the Quebec Integrated Chronic Disease Surveillance System (QICDSS), covering up to 98% of the population in the province of Quebec and containing data for over 20,000 suicides between 2002 and 2019. We employed a case-control study design. Individuals were considered cases if they were aged 15+ and had died from suicide between January 1st, 2002, and December 31st, 2019 (n = 18339). Controls were a random sample of 1% of the Quebec population aged 15+ of each year, who were alive on December 31st of each year, from 2002 to 2019 (n = 1,307,370). We included 103 features, including individual, programmatic, systemic, and community factors, measured up to five years prior to the suicide events. We trained and then validated the sex-specific predictive risk model using supervised ML algorithms, including Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost) and Multilayer perceptron (MLP). We computed operating characteristics, including sensitivity, specificity, and Positive Predictive Value (PPV). We then generated receiver operating characteristic (ROC) curves to predict suicides and calibration measures. For interpretability, Shapley Additive Explanations (SHAP) was used with the global explanation to determine how much the input features contribute to the models’ output and the largest absolute coefficients. The best sensitivity was 0.38 with logistic regression for males and 0.47 with MLP for females; the XGBoost Classifier with 0.25 for males and 0.19 for females had the best precision (PPV). This study demonstrated the useful potential of explainable AI models as tools for decision-making and population-level suicide prevention actions. The ML models included individual, programmatic, systemic, and community levels variables available routinely to decision makers and planners in a public managed care system. Caution shall be exercised in the interpretation of variables associated in a predictive model since they are not causal, and other designs are required to establish the value of individual treatments. The next steps are to produce an intuitive user interface for decision makers, planners and other stakeholders like clinicians or representatives of families and people with live experience of suicidal behaviors or death by suicide. For example, how variations in the quality of local area primary care programs for depression or substance use disorders or increased in regional mental health and addiction budgets would lower suicide rates.
Teacher-Student Curriculum Learning (TSCL) is a curriculum learning framework that draws inspiration from human cultural transmission and le… (see more)arning. It involves a teacher algorithm shaping the learning process of a learner algorithm by exposing it to controlled experiences. Despite its success, understanding the conditions under which TSCL is effective remains challenging. In this paper, we propose a data-centric perspective to analyze the underlying mechanics of the teacher-student interactions in TSCL. We leverage cooperative game theory to describe how the composition of the set of experiences presented by the teacher to the learner, as well as their order, influences the performance of the curriculum that is found by TSCL approaches. To do so, we demonstrate that for every TSCL problem, an equivalent cooperative game exists, and several key components of the TSCL framework can be reinterpreted using game-theoretic principles. Through experiments covering supervised learning, reinforcement learning, and classical games, we estimate the cooperative values of experiences and use value-proportional curriculum mechanisms to construct curricula, even in cases where TSCL struggles. The framework and experimental setup we present in this work represents a novel foundation for a deeper exploration of TSCL, shedding light on its underlying mechanisms and providing insights into its broader applicability in machine learning.
