This paper presents our experiments to quantify the manifolds learned by ML models (in our experiment, we use a GAN model) as they train. We compare the manifolds learned at each epoch to the real manifolds representing the real data. To quantify a manifold, we study the intrinsic dimensions and topological features of the manifold learned by the ML model, how these metrics change as we continue to train the model, and whether these metrics convergence over the course of training to the metrics of the real data manifold.

Effective action abstraction is crucial in tackling challenges associated with large action spaces in Imperfect Information Extensive-Form Games (IIEFGs). However, due to the vast state space and computational complexity in IIEFGs, existing methods often rely on fixed abstractions, resulting in sub-optimal performance. In response, we introduce RL-CFR, a novel reinforcement learning (RL) approach for dynamic action abstraction. RL-CFR builds upon our innovative Markov Decision Process (MDP) formulation, with states corresponding to public information and actions represented as feature vectors indicating specific action abstractions. The reward is defined as the expected payoff difference between the selected and default action abstractions. RL-CFR constructs a game tree with RL-guided action abstractions and utilizes counterfactual regret minimization (CFR) for strategy derivation. Impressively, it can be trained from scratch, achieving higher expected payoff without increased CFR solving time. In experiments on Heads-up No-limit Texas Hold'em, RL-CFR outperforms ReBeL's replication and Slumbot, demonstrating significant win-rate margins of $64\pm 11$ and $84\pm 17$ mbb/hand, respectively.

Tools are essential for large language models (LLMs) to acquire up-to-date information and take consequential actions in external environments. Existing work on tool-augmented LLMs primarily focuses on the broad coverage of tools and the flexibility of adding new tools. However, a critical aspect that has surprisingly been understudied is simply how accurately an LLM uses tools for which it has been trained. We find that existing LLMs, including GPT-4 and open-source LLMs specifically fine-tuned for tool use, only reach a correctness rate in the range of 30% to 60%, far from reliable use in practice. We propose a biologically inspired method for tool-augmented LLMs, simulated trial and error (STE), that orchestrates three key mechanisms for successful tool use behaviors in the biological system: trial and error, imagination, and memory. Specifically, STE leverages an LLM's 'imagination' to simulate plausible scenarios for using a tool, after which the LLM interacts with the tool to learn from its execution feedback. Both short-term and long-term memory are employed to improve the depth and breadth of the exploration, respectively. Comprehensive experiments on ToolBench show that STE substantially improves tool learning for LLMs under both in-context learning and fine-tuning settings, bringing a boost of 46.7% to Mistral-Instruct-7B and enabling it to outperform GPT-4. We also show effective continual learning of tools via a simple experience replay strategy.

Humans perceive the world through multiple senses, enabling them to create a comprehensive representation of their surroundings and to generalize information across domains. For instance, when a textual description of a scene is given, humans can mentally visualize it. In fields like robotics and Reinforcement Learning (RL), agents can also access information about the environment through multiple sensors; yet redundancy and complementarity between sensors is difficult to exploit as a source of robustness (e.g. against sensor failure) or generalization (e.g. transfer across domains). Prior research demonstrated that a robust and flexible multimodal representation can be efficiently constructed based on the cognitive science notion of a 'Global Workspace': a unique representation trained to combine information across modalities, and to broadcast its signal back to each modality. Here, we explore whether such a brain-inspired multimodal representation could be advantageous for RL agents. First, we train a 'Global Workspace' to exploit information collected about the environment via two input modalities (a visual input, or an attribute vector representing the state of the agent and/or its environment). Then, we train a RL agent policy using this frozen Global Workspace. In two distinct environments and tasks, our results reveal the model's ability to perform zero-shot cross-modal transfer between input modalities, i.e. to apply to image inputs a policy previously trained on attribute vectors (and vice-versa), without additional training or fine-tuning. Variants and ablations of the full Global Workspace (including a CLIP-like multimodal representation trained via contrastive learning) did not display the same generalization abilities.

Graph learning plays a pivotal role and has gained significant attention in various application scenarios, from social network analysis to recommendation systems, for its effectiveness in modeling complex data relations represented by graph structural data. In reality, the real-world graph data typically show dynamics over time, with changing node attributes and edge structure, leading to the severe graph data distribution shift issue. This issue is compounded by the diverse and complex nature of distribution shifts, which can significantly impact the performance of graph learning methods in degraded generalization and adaptation capabilities, posing a substantial challenge to their effectiveness. In this survey, we provide a comprehensive review and summary of the latest approaches, strategies, and insights that address distribution shifts within the context of graph learning. Concretely, according to the observability of distributions in the inference stage and the availability of sufficient supervision information in the training stage, we categorize existing graph learning methods into several essential scenarios, including graph domain adaptation learning, graph out-of-distribution learning, and graph continual learning. For each scenario, a detailed taxonomy is proposed, with specific descriptions and discussions of existing progress made in distribution-shifted graph learning. Additionally, we discuss the potential applications and future directions for graph learning under distribution shifts with a systematic analysis of the current state in this field. The survey is positioned to provide general guidance for the development of effective graph learning algorithms in handling graph distribution shifts, and to stimulate future research and advancements in this area.

Evolutionary algorithms, such as Differential Evolution, excel in solving real-parameter optimization challenges. However, the effectiveness of a single algorithm varies across different problem instances, necessitating considerable efforts in algorithm selection or configuration. This paper aims to address the limitation by leveraging the complementary strengths of a group of algorithms and dynamically scheduling them throughout the optimization progress for specific problems. We propose a deep reinforcement learning-based dynamic algorithm selection framework to accomplish this task. Our approach models the dynamic algorithm selection a Markov Decision Process, training an agent in a policy gradient manner to select the most suitable algorithm according to the features observed during the optimization process. To empower the agent with the necessary information, our framework incorporates a thoughtful design of landscape and algorithmic features. Meanwhile, we employ a sophisticated deep neural network model to infer the optimal action, ensuring informed algorithm selections. Additionally, an algorithm context restoration mechanism is embedded to facilitate smooth switching among different algorithms. These mechanisms together enable our framework to seamlessly select and switch algorithms in a dynamic online fashion. Notably, the proposed framework is simple and generic, offering potential improvements across a broad spectrum of evolutionary algorithms. As a proof-of-principle study, we apply this framework to a group of Differential Evolution algorithms. The experimental results showcase the remarkable effectiveness of the proposed framework, not only enhancing the overall optimization performance but also demonstrating favorable generalization ability across different problem classes.

At first glance, if you squint, you might think it was a photograph: a couple nuzzling together in the front row of a concert hall, in a Facebook advertisement for the Queensland Symphony Orchestra (QSO). But look again and you'll see why it's caused a stir among creative workers and the union representing them. The couple's tangled fingers are both too large and too many; there's a strange sheen making them look more like wax dolls; and then there's the clothes: she in a tulle gown encrusted with jewels, he in a tuxedo – and, simultaneously, a tulle gown encrusted with jewels. Also: she has a large cube on her lap. "Want to do something different this Saturday? Come see an orchestra play," reads the ad.

We propose a method to teach multiple large language models (LLM) to collaborate by interleaving their generations at the token level. We model the decision of which LLM generates the next token as a latent variable. By optimizing the marginal likelihood of a training set under our latent variable model, the base LLM automatically learns when to generate itself and when to call on one of the ``assistant'' language models to generate, all without direct supervision. Token-level collaboration during decoding allows for a fusion of each model's expertise in a manner tailored to the specific task at hand. Our collaborative decoding is especially useful in cross-domain settings where a generalist base LLM learns to invoke domain expert models. On instruction-following, domain-specific QA, and reasoning tasks, we show that the performance of the joint system exceeds that of the individual models. Through qualitative analysis of the learned latent decisions, we show models trained with our method exhibit several interesting collaboration patterns, e.g., template-filling. Our code is available at https://github.com/clinicalml/co-llm.

With the advancement of large language models, language-based forecasting has recently emerged as an innovative approach for predicting human mobility patterns. The core idea is to use prompts to transform the raw mobility data given as numerical values into natural language sentences so that the language models can be leveraged to generate the description for future observations. However, previous studies have only employed fixed and manually designed templates to transform numerical values into sentences. Since the forecasting performance of language models heavily relies on prompts, using fixed templates for prompting may limit the forecasting capability of language models. In this paper, we propose a novel framework for prompt mining in language-based mobility forecasting, aiming to explore diverse prompt design strategies. Specifically, the framework includes a prompt generation stage based on the information entropy of prompts and a prompt refinement stage to integrate mechanisms such as the chain of thought. Experimental results on real-world large-scale data demonstrate the superiority of generated prompts from our prompt mining pipeline. Additionally, the comparison of different prompt variants shows that the proposed prompt refinement process is effective. Our study presents a promising direction for further advancing language-based mobility forecasting.

In this paper, we study a defense against poisoned encoders in SSL called distillation, which is a defense used in supervised learning originally. Distillation aims to distill knowledge from a given model (a.k.a the teacher net) and transfer it to another (a.k.a the student net). Now, we use it to distill benign knowledge from poisoned pre-trained encoders and transfer it to a new encoder, resulting in a clean pre-trained encoder. In particular, we conduct an empirical study on the effectiveness and performance of distillation against poisoned encoders. Using two state-of-the-art backdoor attacks against pre-trained image encoders and four commonly used image classification datasets, our experimental results show that distillation can reduce attack success rate from 80.87% to 27.51% while suffering a 6.35% loss in accuracy. Moreover, we investigate the impact of three core components of distillation on performance: teacher net, student net, and distillation loss. By comparing 4 different teacher nets, 3 student nets, and 6 distillation losses, we find that fine-tuned teacher nets, warm-up-training-based student nets, and attention-based distillation loss perform best, respectively.