Manzil Zaheer
Research Areas
Authored Publications
Sort By
Preview abstract
Many modern high-performing machine learning models such as GPT-3 primarily rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference.
Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Remarkably, retrieval-based methods have enjoyed success on a wide range of problems, ranging from standard natural language processing and vision tasks to protein folding, as demonstrated by many recent efforts, including WebGPT and AlphaFold. Despite a growing literature showcasing the promise of these models, the theoretical underpinning for such models remains underexplored. In this paper, we present a formal treatment of retrieval-based models to characterize their generalization ability. In particular, we focus on two classes of retrieval-based classification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task.
View details
Teacher Guided Training: An Efficient Framework for Knowledge Transfer
Chong You
Himanshu Jain
Rob Fergus
International Conference on Learning Representations(2023) (to appear)
Preview abstract
The remarkable performance gains realized by large pretrained models, e.g., GPT-3, hinge on the massive amounts of data they are exposed to during training. Analogously, distilling such large models to compact models for efficient deployment also necessitates a large amount of (labeled or unlabeled) training data. In this paper, we devise teacher-guided training (TGT) framework for training a high-quality compact model that leverages the knowledge acquired by pre-trained \emph{generative} models while obviating the need to go through a large volume of data. TGT exploits the fact that the teacher has acquired a good representation of the underlying data domain, which typically corresponds to a much lower dimensional manifold than the ambient space. Furthermore, we can use the teacher to explore the instance space more efficiently through sampling or gradient-based methods; thus, making TGT especially attractive for limited data or long-tail settings. We formally capture this benefit of proposed data-domain exploration in our generalization bounds. Among our empirical evaluations, we find that TGT can improve accuracy on ImageNet-LT by 10% compared to natural baseline and match accuracy on sentiment analysis on Amazon reviews without the need for pretraining.
View details
A Fourier Approach to Mixture Learning
Mingda Qiao
Guru Prashanth Guruganesh
Conference on Neural Information Processing Systems(2022) (to appear)
Preview abstract
We revisit the problem of learning mixtures of spherical Gaussians. Given samples from mixture $\frac{1}{k}\sum_{j=1}^{k}\N(\mu_j, I_d)$, the goal is to estimate the means $\mu_1, \mu_2, \ldots, \mu_k \in \R^d$ up to a small error. The hardness of this learning problem can be measured by the \emph{separation} $\Delta$ defined as the minimum distance between all pairs of means. Regev and Vijayaraghavan (2017) showed that with $\Delta = \Omega(\sqrt{\log k})$ separation, the means can be learned using $\poly(k, d)$ samples, whereas super-polynomially many samples are required if $\Delta = o(\sqrt{\log k})$ and $d = \Omega(\log k)$. This leaves open the low-dimensional regime where $d = o(\log k)$.
In this work, we give an algorithm that efficiently learns the means in $d = O(\log k/\log\log k)$ dimensions under separation $d/\sqrt{\log k}$ (modulo doubly logarithmic factors). This separation is strictly smaller than $\sqrt{\log k}$, and is also shown to be necessary. Along with the results of Regev and Vijayaraghavan (2017), our work almost pins down the critical separation threshold at which efficient parameter learning becomes possible for spherical Gaussian mixtures. This was previously open even in one dimension. More generally, our algorithm runs in time $\poly(k)\cdot f(d, \Delta, \eps)$, and is thus fixed-parameter tractable in parameters $d$, $\Delta$ and $\eps$.
Our approach is based on estimating the Fourier transform of the mixture at carefully chosen frequencies, and both the algorithm and its analysis are simple and elementary. Our positive results can be easily extended to learning mixtures of non-Gaussian distributions, under a mild condition on the Fourier spectrum of the distribution.
View details
Thompson Sampling with a Mixture Prior
Joey Hong
Branislav Kveton
Mohammad Ghavamzadeh
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics (AI-Stats-22)(2022), pp. 7565-7586
Preview abstract
We consider posterior sampling in online decision-making problems where the uncertain environment is sampled from a mixture distribution. We incorporate this structure in a natural way by initializing a Thompson sampling algorithm with a mixture prior. We provide a novel, general outline for analyzing the regret of Thompson sampling with a mixture prior. We also use this to derive Bayes regret bounds in both a linear bandit and tabular MDP settings. The regret bounds depend on the confidence widths of each component of the mixture prior, and converge to solely identifying the correct component when confidence widths are small. Finally, we demonstrate the empirical effectiveness of our proposed algorithm in a synthetic and real-world bandit problem involving multi-task image classification.
View details
A Context Integrated Transformer-based Neural Network for Auction Design
Zhijian Duan
Jingwu Tang
Yutong Yin
Xiang Yan
Xiaotie Deng
The Thirty-ninth International Conference on Machine Learning (ICML'22)(2022)
Preview abstract
One of the central problems in auction design is to develop an incentive compatible mechanism that maximizes the expected revenue. While theoretical approaches have encountered bottlenecks for multi-item auctions, recently there are many progresses of finding optimal auction through deep learning.
However, such works either focus on a fixed set of bidders and items, or restrict the auction to be symmetric. In this work, we overcome this limitation by factoring \emph{public} contextual information of bidders and items into deep learning framework. We propose $\mathtt{CITransNet}$, a context integrated transformer-based neural network for contextual auction design, which maintains permutation-equivariance over bids while being able to handle asymmetric contextual information in auctions. We show by extensive experiments that $\mathtt{CITransNet}$ can recover the known optimal analytical solutions, obtain novel mechanisms for complex multi-item auctions, and generalize to settings different from training set.
View details
Compositional Generalization and Decomposition in Neural Program Synthesis
Joey Hong
Deep Learning for Code (DL4C) Workshop at ICLR'22(2022)
Preview abstract
When writing programs, people have the ability to tackle a new complex task by decomposing it into smaller and more familiar subtasks. While it is difficult to measure whether neural program synthesis methods have similar capabilities, what we can measure is whether they compositionally generalize, that is, whether a model that has been trained on the simpler subtasks is subsequently able to solve more complex tasks. In this paper, we focus on measuring the ability of learned program synthesizers to compositionally generalize. We first characterize several different axes along which program synthesis methods would be desired to generalize, e.g., length generalization, or the ability to combine known subroutines in new ways that do not occur in the training data. Based on this characterization, we introduce a benchmark suite of tasks to assess these abilities based on two popular existing datasets, SCAN and RobustFill. Finally, we make first attempts to improve the compositional generalization ability of Transformer models along these axes through novel attention mechanisms that draw inspiration from a human-like decomposition strategy. Empirically, we find our modified Transformer models generally perform better than natural baselines, but the tasks remain challenging.
View details
Meta-Thompson Sampling
Branislav Kveton
Michael Konobeev
Martin Mladenov
Proceedings of the 38th International Conference on Machine Learning (ICML 2021), pp. 5884-5893
Preview abstract
Efficient exploration in multi-armed bandits is a fundamental online learning problem. In this work, we propose a variant of Thompson sampling that learns to explore over time by interacting with problem instances sampled from an unknown prior distribution. This algorithm meta-learns the prior and therefore we call it Meta-TS. We propose efficient implementations of Meta-TS and analyze it in Gaussian bandits. Our analysis captures the improvement due to learning the prior and is of a broader interest, because we derive the first prior-dependent upper bound on the Bayes regret. Our regret bound is complemented by empirical evaluation, which shows that Meta-TS quickly adapts to the unknown prior.
View details
Scalable Hierarchical Agglomerative Clustering
Nick Monath
Guru Prashanth Guruganesh
Andrew McCallum
Gokhan Mergen
Mert Terzihan
Bryon Tjanaka
Yuchen Wu
Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining(2021), 1245–1255
Preview abstract
The applicability of agglomerative clustering, for inferring both hierarchical and flat clustering, is limited by its scalability. Existing scalable hierarchical clustering methods sacrifice quality for speed and often lead to over-merging of clusters. In this paper, we present a scalable, agglomerative method for hierarchical clustering that does not sacrifice quality and scales to billions of data points. We perform a detailed theoretical analysis, showing that under mild separability conditions our algorithm can not only recover the optimal flat partition but also provide a two-approximation to non-parametric DP-Means objective [32]. This introduces a novel application of hierarchical clustering as an approximation algorithm for the non-parametric clustering objective. We additionally relate our algorithm to the classic hierarchical agglomerative clustering method. We perform extensive empirical experiments in both hierarchical and flat clustering settings and show that our proposed approach achieves state-of-the-art results on publicly available clustering benchmarks. Finally, we demonstrate our method’s scalability by applying it to a dataset of 30 billion queries. Human evaluation of the discovered clusters show that our method finds better quality of clusters than the current state-of-the-art.
View details
Latent Programmer: Discrete Latent Codes for Program Synthesis
Joey Hong
David Martin Dohan
Rishabh Singh
International Conference on Machine Learning (ICML)(2021)
Preview abstract
In many sequence learning tasks, such as program synthesis and document summarization, a key problem is searching over a large space of possible output sequences.
We propose to learn representations of the outputs that is specifically meant for search: rich enough to specify the desired output but compact enough to make search more efficient.
An appealing realization of such representation are discrete latent codes, as this naturally allows sophisticated combinatorial search strategies.
The latent codes are learned using a self-supervised learning principle, in which first a discrete autoencoder is trained on the output sequences, and then the resulting latent codes are used as intermediate targets for the end-to-end sequence prediction task.
Based on these insights, we introduce the Latent Programmer, a program synthesis method that first predicts a discrete latent codes from input/output examples, and then generates the program in the target language.
We evaluate the Latent Programmer on two domains: synthesis of string transformation programs, and generation of programs from natural language descriptions.
We demonstrate that the discrete latent representation significantly improves synthesis accuracy.
View details
Non-Stationary Off-policy Optimization
Joey Hong
Branislav Kveton
International Conference on Artificial Intelligence and Statistics (AISTATS)(2021)
Preview abstract
Off-policy learning is a framework for estimating the value of and optimizing policies offline from logged data without deploying them. Real-world environments are nonstationary, and the optimized policies should be able to adapt to these changes. To address this challenge, we study the novel problem of off-policy optimization in piecewise-stationary environments. Our key idea is to use a change-point detector to partition the logged data into categorical latent states, then find a near-optimal policy conditioned on latent state. We derive high-probability bounds on our off-policy estimates and optimization. Furthermore, we also propose a practical approach to deploy our policy online and evaluate our approach comprehensively on a real-world clickstream dataset.
View details
