Specifying objective functions for machine learning systems is challenging, and misspecified objectives can be hacked [1, 2] or incentivise degenerate behavior [3, 4, 5]. Techniques such as reinforcement learning from human feedback (RLHF) enable ML systems to instead learn appropriate objectives from human feedback [6, 7, 8]. These techniques are widely used to finetune large language models [9, 10, 11, 12] and to train reinforcement learning agents to perform complex maneuvers in continuous control environments [6, 7]. However, while RLHF is relied upon to ensure that these systems are safe, helpful, and harmless [13], it still faces many limitations and unsolved challenges [14]. In particular, RLHF systems typically rely on the assumption that all feedback comes from a single human teacher, despite gathering feedback from a range of teachers with varying levels of rationality and expertise. For example, Stiennon et al. [8], Bai et al. [13] and Ouyang et al. [15] assume that all feedback comes from a single teacher, but find that annotators and researchers actually disagree 23% to 37% of the time. Reward learning has been shown to be highly sensitive to incorrect assumptions about the process that generates feedback [16, 17, 18, 19], so this single-teacher assumption exposes these systems to dangerous failures [20]. Ideally, RLHF systems should consider the differences between each teacher to improve their safety and reliability. To leverage multiple teachers in RLHF, we introduce a novel problem called a Hidden Utility Bandit (HUB), illustrated in Figure 1.

Exposing meaningful and interpretable neural interactions is critical to understanding neural circuits. Inferred neural interactions from neural signals primarily reflect functional interactions. In a long experiment, subject animals may experience different stages defined by the experiment, stimuli, or behavioral states, and hence functional interactions can change over time. To model dynamically changing functional interactions, prior work employs state-switching generalized linear models with hidden Markov models (i.e., HMM-GLMs). However, we argue they lack biological plausibility, as functional interactions are shaped and confined by the underlying anatomical connectome. Here, we propose a novel prior-informed state-switching GLM. We introduce both a Gaussian prior and a one-hot prior over the GLM in each state. The priors are learnable. We will show that the learned prior should capture the state-constant interaction, shedding light on the underlying anatomical connectome and revealing more likely physical neuron interactions. The state-dependent interaction modeled by each GLM offers traceability to capture functional variations across multiple brain states. Our methods effectively recover true interaction structures in simulated data, achieve the highest predictive likelihood with real neural datasets, and render interaction structures and hidden states more interpretable when applied to real neural data.

Community Question-Answering (CQA) portals serve as a valuable tool for helping users within an organization. However, making them accessible to non-English-speaking users continues to be a challenge. Translating questions can broaden the community's reach, benefiting individuals with similar inquiries in various languages. Translating questions using Neural Machine Translation (NMT) poses more challenges, especially in noisy environments, where the grammatical correctness of the questions is not monitored. These questions may be phrased as statements by non-native speakers, with incorrect subject-verb order and sometimes even missing question marks. Creating a synthetic parallel corpus from such data is also difficult due to its noisy nature. To address this issue, we propose a training methodology that fine-tunes the NMT system only using source-side data. Our approach balances adequacy and fluency by utilizing a loss function that combines BERTScore and Masked Language Model (MLM) Score. Our method surpasses the conventional Maximum Likelihood Estimation (MLE) based fine-tuning approach, which relies on synthetic target data, by achieving a 1.9 BLEU score improvement. Our model exhibits robustness while we add noise to our baseline, and still achieve 1.1 BLEU improvement and large improvements on TER and BLEURT metrics. Our proposed methodology is model-agnostic and is only necessary during the training phase. We make the codes and datasets publicly available at \url{https://www.iitp.ac.in/~ai-nlp-ml/resources.html#DomainAdapt} for facilitating further research.

In a previous work we introduced, in the context of gravitational wave science, an initial study on an automated domain-decomposition approach for reduced basis through hp-greedy refinement. The approach constructs local reduced bases of lower dimensionality than global ones, with the same or higher accuracy. These ``light'' local bases should imply both faster evaluations when predicting new waveforms and faster data analysis, in particular faster statistical inference (the forward and inverse problems, respectively). In this approach, however, we have previously found important dependence on several hyperparameters, which do not appear in global reduced basis. This naturally leads to the problem of hyperparameter optimization (HPO), which is the subject of this paper. We tackle the problem through a Bayesian optimization, and show its superiority when compared to grid or random searches. We find that for gravitational waves from the collision of two spinning but non-precessing black holes, for the same accuracy, local hp-greedy reduced bases with HPO have a lower dimensionality of up to $4 \times$ for the cases here studied, depending on the desired accuracy. This factor should directly translate in a parameter estimation speedup, for instance. Such acceleration might help in the near real-time requirements for electromagnetic counterparts of gravitational waves from compact binary coalescences. In addition, we find that the Bayesian approach used in this paper for HPO is two orders of magnitude faster than, for example, a grid search, with about a $100 \times$ acceleration. The code developed for this project is available as open source from public repositories.

Data is the lifeblood of machine learning, yet for many firms, obtaining datasets of sufficient quality remains a challenge, with them being naturally distributed amongst owners with heterogeneous characteristics (e.g., privacy preferences). This has motivated several developments in the field of collaborative analytics, also known as federated learning (Figure 1a), where models are trained on local servers without the need for data centralization, thereby preserving privacy and distributing the computational burden (Kairouz et al., 2019). However, this framework provides only an incentive-free means for data sharing, relying on the critical assumption that owners are willing to collaborate (i.e., by sharing their private information) altruistically. This rather strong assumption may be violated if owners are competitors in a downstream market environment (Gal-Or, 1985). Consequently, a fruitful area of research has emerged that proposes to instead commoditize data within a market-based framework, where compensation (e.g., remuneration) can be used as an incentive for collaboration (Bergemann and Bonatti, 2019).

Advances in single-cell omics allow for unprecedented insights into the transcription profiles of individual cells. When combined with large-scale perturbation screens, through which specific biological mechanisms can be targeted, these technologies allow for measuring the effect of targeted perturbations on the whole transcriptome. These advances provide an opportunity to better understand the causative role of genes in complex biological processes such as gene regulation, disease progression or cellular development. However, the high-dimensional nature of the data, coupled with the intricate complexity of biological systems renders this task nontrivial. Within the machine learning community, there has been a recent increase of interest in causality, with a focus on adapting established causal techniques and algorithms to handle high-dimensional data. In this perspective, we delineate the application of these methodologies within the realm of single-cell genomics and their challenges. We first present the model that underlies most of current causal approaches to single-cell biology and discuss and challenge the assumptions it entails from the biological point of view. We then identify open problems in the application of causal approaches to single-cell data: generalising to unseen environments, learning interpretable models, and learning causal models of dynamics. For each problem, we discuss how various research directions - including the development of computational approaches and the adaptation of experimental protocols - may offer ways forward, or on the contrary pose some difficulties. With the advent of single cell atlases and increasing perturbation data, we expect causal models to become a crucial tool for informed experimental design.

The posterior in probabilistic programs with stochastic support decomposes as a weighted sum of the local posterior distributions associated with each possible program path. We show that making predictions with this full posterior implicitly performs a Bayesian model averaging (BMA) over paths. This is potentially problematic, as model misspecification can cause the BMA weights to prematurely collapse onto a single path, leading to sub-optimal predictions in turn. To remedy this issue, we propose alternative mechanisms for path weighting: one based on stacking and one based on ideas from PAC-Bayes. We show how both can be implemented as a cheap post-processing step on top of existing inference engines. In our experiments, we find them to be more robust and lead to better predictions compared to the default BMA weights.

Solving complex classification tasks using deep neural networks typically requires large amounts of annotated data. However, corresponding class labels are noisy when provided by error-prone annotators, e.g., crowdworkers. Training standard deep neural networks leads to subpar performances in such multi-annotator supervised learning settings. We address this issue by presenting a probabilistic training framework named multi-annotator deep learning (MaDL). A downstream ground truth and an annotator performance model are jointly trained in an end-to-end learning approach. The ground truth model learns to predict instances' true class labels, while the annotator performance model infers probabilistic estimates of annotators' performances. A modular network architecture enables us to make varying assumptions regarding annotators' performances, e.g., an optional class or instance dependency. Further, we learn annotator embeddings to estimate annotators' densities within a latent space as proxies of their potentially correlated annotations. Together with a weighted loss function, we improve the learning from correlated annotation patterns. In a comprehensive evaluation, we examine three research questions about multi-annotator supervised learning. Our findings show MaDL's state-of-the-art performance and robustness against many correlated, spamming annotators.

Can non-programmers annotate natural language utterances with complex programs that represent their meaning? We introduce APEL, a framework in which non-programmers select among candidate programs generated by a seed semantic parser (e.g., Codex). Since they cannot understand the candidate programs, we ask them to select indirectly by examining the programs' input-ouput examples. For each utterance, APEL actively searches for a simple input on which the candidate programs tend to produce different outputs. It then asks the non-programmers only to choose the appropriate output, thus allowing us to infer which program is correct and could be used to fine-tune the parser. As a first case study, we recruited human non-programmers to use APEL to re-annotate SPIDER, a text-to-SQL dataset. Our approach achieved the same annotation accuracy as the original expert annotators (75%) and exposed many subtle errors in the original annotations.

We propose a new regret minimization algorithm for episodic sparse linear Markov decision process (SMDP) where the state-transition distribution is a linear function of observed features. The only previously known algorithm for SMDP requires the knowledge of the sparsity parameter and oracle access to an unknown policy. We overcome these limitations by combining the doubly robust method that allows one to use feature vectors of \emph{all} actions with a novel analysis technique that enables the algorithm to use data from all periods in all episodes. The regret of the proposed algorithm is $\tilde{O}(\sigma^{-1}_{\min} s_{\star} H \sqrt{N})$, where $\sigma_{\min}$ denotes the restrictive the minimum eigenvalue of the average Gram matrix of feature vectors, $s_\star$ is the sparsity parameter, $H$ is the length of an episode, and $N$ is the number of rounds. We provide a lower regret bound that matches the upper bound up to logarithmic factors on a newly identified subclass of SMDPs. Our numerical experiments support our theoretical results and demonstrate the superior performance of our algorithm.