We study multiclass PAC learning with bandit feedback, where inputs are classified into one of $K$ possible labels and feedback is limited to whether or not the predicted labels are correct. Our main contribution is in designing a novel learning algorithm for the agnostic $(\varepsilon,\delta)$-PAC version of the problem, with sample complexity of $O\big( (\operatorname{poly}(K) + 1 / \varepsilon^2) \log (|H| / \delta) \big)$ for any finite hypothesis class $H$. In terms of the leading dependence on $\varepsilon$, this improves upon existing bounds for the problem, that are of the form $O(K/\varepsilon^2)$. We also provide an extension of this result to general classes and establish similar sample complexity bounds in which $\log |H|$ is replaced by the Natarajan dimension. This matches the optimal rate in the full-information version of the problem and resolves an open question studied by Daniely, Sabato, Ben-David, and Shalev-Shwartz (2011) who demonstrated that the multiplicative price of bandit feedback in realizable PAC learning is $\Theta(K)$. We complement this by revealing a stark contrast with the agnostic case, where the price of bandit feedback is only $O(1)$ as $\varepsilon \to 0$. Our algorithm utilizes a stochastic optimization technique to minimize a log-barrier potential based on Frank-Wolfe updates for computing a low-variance exploration distribution over the hypotheses, and is made computationally efficient provided access to an ERM oracle over $H$.

Solving mathematical problems requires advanced reasoning abilities and presents notable challenges for large language models. Previous works usually synthesize data from proprietary models to augment existing datasets, followed by instruction tuning to achieve top-tier results. However, our analysis of these datasets reveals severe biases towards easy queries, with frequent failures to generate any correct response for the most challenging queries. Hypothesizing that difficult queries are crucial to learn complex reasoning, we propose Difficulty-Aware Rejection Tuning (DART), a method that allocates difficult queries more trials during the synthesis phase, enabling more extensive training on difficult samples. Utilizing DART, we have created new datasets for mathematical problem-solving that focus more on difficult queries and are substantially smaller than previous ones. Remarkably, our synthesis process solely relies on a 7B-sized open-weight model, without reliance on the commonly used proprietary GPT-4. We fine-tune various base models on our datasets ranging from 7B to 70B in size, resulting in a series of strong models called DART-Math. In comprehensive in-domain and out-of-domain evaluation on 6 mathematical benchmarks, DART-Math outperforms vanilla rejection tuning significantly, being superior or comparable to previous arts, despite using much smaller datasets and no proprietary models.

The critical inquiry pervading the realm of Philosophy, and perhaps extending its influence across all Humanities disciplines, revolves around the intricacies of morality and normativity. Surprisingly, in recent years, this thematic thread has woven its way into an unexpected domain, one not conventionally associated with pondering "what ought to be": the field of artificial intelligence (AI) research. Central to morality and AI, we find "alignment", a problem related to the challenges of expressing human goals and values in a manner that artificial systems can follow without leading to unwanted adversarial effects. More explicitly and with our current paradigm of AI development in mind, we can think of alignment as teaching human values to non-anthropomorphic entities trained through opaque, gradient-based learning techniques. This work addresses alignment as a technical-philosophical problem that requires solid philosophical foundations and practical implementations that bring normative theory to AI system development. To accomplish this, we propose two sets of necessary and sufficient conditions that, we argue, should be considered in any alignment process. While necessary conditions serve as metaphysical and metaethical roots that pertain to the permissibility of alignment, sufficient conditions establish a blueprint for aligning AI systems under a learning-based paradigm. After laying such foundations, we present implementations of this approach by using state-of-the-art techniques and methods for aligning general-purpose language systems. We call this framework Dynamic Normativity. Its central thesis is that any alignment process under a learning paradigm that cannot fulfill its necessary and sufficient conditions will fail in producing aligned systems.

Federated learning is a promising privacy-preserving paradigm for distributed machine learning. In this context, there is sometimes a need for a specialized process called machine unlearning, which is required when the effect of some specific training samples needs to be removed from a learning model due to privacy, security, usability, and/or legislative factors. However, problems arise when current centralized unlearning methods are applied to existing federated learning, in which the server aims to remove all information about a class from the global model. Centralized unlearning usually focuses on simple models or is premised on the ability to access all training data at a central node. However, training data cannot be accessed on the server under the federated learning paradigm, conflicting with the requirements of the centralized unlearning process. Additionally, there are high computation and communication costs associated with accessing clients' data, especially in scenarios involving numerous clients or complex global models. To address these concerns, we propose a more effective and efficient federated unlearning scheme based on the concept of model explanation. Model explanation involves understanding deep networks and individual channel importance, so that this understanding can be used to determine which model channels are critical for classes that need to be unlearned. We select the most influential channels within an already-trained model for the data that need to be unlearned and fine-tune only influential channels to remove the contribution made by those data. In this way, we can simultaneously avoid huge consumption costs and ensure that the unlearned model maintains good performance. Experiments with different training models on various datasets demonstrate the effectiveness of the proposed approach.

Large Language Models (LLMs) possess the potential to exert substantial influence on public perceptions and interactions with information. This raises concerns about the societal impact that could arise if the ideologies within these models can be easily manipulated. In this work, we investigate how effectively LLMs can learn and generalize ideological biases from their instruction-tuning data. Our findings reveal a concerning vulnerability: exposure to only a small amount of ideologically driven samples significantly alters the ideology of LLMs. Notably, LLMs demonstrate a startling ability to absorb ideology from one topic and generalize it to even unrelated ones. The ease with which LLMs' ideologies can be skewed underscores the risks associated with intentionally poisoned training data by malicious actors or inadvertently introduced biases by data annotators. It also emphasizes the imperative for robust safeguards to mitigate the influence of ideological manipulations on LLMs.

We assess the capability of GPT-3.5, LLaMA-2 iterative editing to ensure that the revised texts meet the 70B, and Mixtral 8x7B, to generate content at various readability desired difficulty criteria. This readability assessment is based on levels through zero-shot and few-shot prompting. Evaluating 100 various linguistic features, with sentence length and word frequency processed educational materials reveals that few-shot prompting identified as key factors in previous studies [11]. Although this significantly improves performance in readability manipulation and process appears straightforward, accurately adjusting these elements information preservation. LLaMA-2 70B performs better in achieving to achieve the target reading difficulty is challenging. This the desired difficulty range, while GPT-3.5 maintains original task becomes even more complex for young learners, where factors meaning. However, manual inspection highlights concerns such such as decodability [19], information load [15], and other elements as misinformation introduction and inconsistent edit distribution.

Detecting evidence within the context is a key step in the process of reasoning task. Evaluating and enhancing the capabilities of LLMs in evidence detection will strengthen context-based reasoning performance. This paper proposes a benchmark called DetectBench for verifying the ability to detect and piece together implicit evidence within a long context. DetectBench contains 3,928 multiple-choice questions, with an average of 994 tokens per question. Each question contains an average of 4.55 pieces of implicit evidence, and solving the problem typically requires 7.62 logical jumps to find the correct answer. To enhance the performance of LLMs in evidence detection, this paper proposes Detective Reasoning Prompt and Finetune. Experiments demonstrate that the existing LLMs' abilities to detect evidence in long contexts are far inferior to humans. However, the Detective Reasoning Prompt effectively enhances the capability of powerful LLMs in evidence detection, while the Finetuning method shows significant effects in enhancing the performance of weaker LLMs. Moreover, when the abilities of LLMs in evidence detection are improved, their final reasoning performance is also enhanced accordingly.

We study the evolution of opinions inside a population of interacting large language models (LLMs). Every LLM needs to decide how much funding to allocate to an item with three initial possibilities: full, partial, or no funding. We identify biases that drive the exchange of opinions based on the LLM's tendency to (i) find consensus with the other LLM's opinion, (ii) display caution when specifying funding, and (iii) consider ethical concerns in its opinion. We find these biases are affected by the perceived absence of compelling reasons for opinion change, the perceived willingness to engage in discussion, and the distribution of allocation values. Moreover, tensions among biases can lead to the survival of funding for items with negative connotations. We also find that the final distribution of full, partial, and no funding opinions is more diverse when an LLM freely forms its opinion after an interaction than when its opinion is a multiple-choice selection among the three allocation options. In the latter case, consensus or polarization is generally attained. When agents are aware of past opinions, they seek to maintain consistency with them, and more diverse updating rules emerge. Our study is performed using a Llama 3 LLM.

They lack a systematic review ARGE language models (LLMs) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] have posed a groundbreaking impact on artificial and discussion on preference learning, especially preference intelligence (AI), transforming the opinions of people on the modeling methods, which are critical to capturing human potential of AI systems for understanding and applying human intentions for LM alignment [13]. To further explore effective languages. These neural network language models with largescale preference learning approaches for better LLM alignment, parameters (mainly over 10 billion) are initially pretrained we present a comprehensive review of human preference on large corpora collected from a wide range of learning methods applicable to language models, examining sources, a remarkable part of which is on the Internet [11]. After LLM alignment methods from the perspective of preference pre-training by imitating how humans use natural languages learning. By analyzing a wide range of alignment approaches in the text data, the foundation LLMs acquire strong and within the preference learning framework, we outline the general language skills [1, 12]. On the other hand, foundation holistic picture of introducing human preference into LLMs, LLMs are observed to have difficulty in understanding or enabling insights to be drawn from every aspect of human responding to diverse human instruction appropriately [13], preference learning for various domains.

Instruction-following large language models (LLMs), such as ChatGPT, have become increasingly popular with the general audience, many of whom are incorporating them into their daily routines. However, these LLMs inadvertently disclose personal or copyrighted information, which calls for a machine unlearning method to remove selective knowledge. Previous attempts sought to forget the link between the target information and its associated entities, but it rather led to generating undesirable responses about the target, compromising the end-user experience. In this work, we propose SNAP, an innovative framework designed to selectively unlearn information by 1) training an LLM with negative instructions to generate obliterated responses, 2) augmenting hard positives to retain the original LLM performance, and 3) applying the novel Wasserstein regularization to ensure adequate deviation from the initial weights of the LLM. We evaluate our framework on various NLP benchmarks and demonstrate that our approach retains the original LLM capabilities, while successfully unlearning the specified information.