Author name cluster

Jilin Chen

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

ICLR Conference 2024 Conference Paper

Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering

Han Zhou 0010
Xingchen Wan
Lev Proleev
Diana Mincu
Jilin Chen
Katherine A. Heller
Subhrajit Roy

Prompting and in-context learning (ICL) have become efficient learning paradigms for large language models (LLMs). However, LLMs suffer from prompt brittleness and various bias factors in the prompt, including but not limited to the formatting, the choice verbalizers, and the ICL examples. To address this problem that results in unexpected performance degradation, calibration methods have been developed to mitigate the effects of these biases while recovering LLM performance. In this work, we first conduct a systematic analysis of the existing calibration methods, where we both provide a unified view and reveal the failure cases. Inspired by these analyses, we propose Batch Calibration (BC), a simple yet intuitive method that controls the contextual bias from the batched input, unifies various prior approaches and effectively addresses the aforementioned issues. BC is zero-shot, inference-only, and incurs negligible additional costs. In the few-shot setup, we further extend BC to allow it to learn the contextual bias from labeled data. We validate the effectiveness of BC with PaLM 2-(S, M, L) and CLIP models and demonstrate state-of-the-art performance over previous calibration baselines across more than 10 natural language understanding and image classification tasks.

Details

TMLR Journal 2024 Journal Article

Break it, Imitate it, Fix it: Robustness by Generating Human-Like Attacks

Aradhana Sinha
Ananth Balashankar
Ahmad Beirami
Thi Avrahami
Jilin Chen
Alex Beutel

Real-world natural language processing systems need to be robust to human adversaries. Collecting examples of human adversaries for training is an effective but expensive solution. On the other hand, training on synthetic attacks with small perturbations---such as word-substitution---does not actually improve robustness to human adversaries. In this paper, we propose an adversarial training framework that uses limited human adversarial examples to generate more useful adversarial examples at scale. We demonstrate the advantages of this system on the ANLI and hate speech detection benchmark datasets---both collected via an iterative, adversarial human-and-model-in-the-loop procedure. Compared to training only on observed human attacks, also training on our synthetic adversarial examples improves model robustness to future rounds. In ANLI, we see accuracy gains on the current set of attacks (44.1\%$\,\to\,$50.1\%) and on two future unseen rounds of human generated attacks (32.5\%$\,\to\,$43.4\%, and 29.4\%$\,\to\,$40.2\%). In hate speech detection, we see AUC gains on current attacks (0.76 $\to$ 0.84) and a future round (0.77 $\to$ 0.79). Attacks from methods that do not learn the distribution of existing human adversaries, meanwhile, degrade robustness.

PDF Details

ICML Conference 2024 Conference Paper

Controlled Decoding from Language Models

Sidharth Mudgal
Jong Lee
Harish Ganapathy
YaGuang Li
Tao Wang
Yanping Huang
Zhifeng Chen
Heng-Tze Cheng

KL-regularized reinforcement learning (RL) is a popular alignment framework to control the language model responses towards high reward outcomes. We pose a tokenwise RL objective and propose a modular solver for it, called controlled decoding (CD). CD exerts control through a separate prefix scorer module, which is trained to learn a value function for the reward. The prefix scorer is used at inference time to control the generation from a frozen base model, provably sampling from a solution to the RL objective. We empirically demonstrate that CD is effective as a control mechanism on popular benchmarks. We also show that prefix scorers for multiple rewards may be combined at inference time, effectively solving a multi-objective RL problem with no additional training. We show that the benefits of applying CD transfer to an unseen base model with no further tuning as well. Finally, we show that CD can be applied in a blockwise decoding fashion at inference-time, essentially bridging the gap between the popular best-of-$K$ strategy and tokenwise control through reinforcement learning. This makes CD a promising approach for alignment of language models.

Details

NeurIPS Conference 2020 Conference Paper

Fairness without Demographics through Adversarially Reweighted Learning

Preethi Lahoti
Alex Beutel
Jilin Chen
Kang Lee
Flavien Prost
Nithum Thain
Xuezhi Wang
Ed Chi

Much of the previous machine learning (ML) fairness literature assumes that protected features such as race and sex are present in the dataset, and relies upon them to mitigate fairness concerns. However, in practice factors like privacy and regulation often preclude the collection of protected features, or their use for training or inference, severely limiting the applicability of traditional fairness research. Therefore, we ask: How can we train a ML model to improve fairness when we do not even know the protected group memberships? In this work we address this problem by proposing Adversarially Reweighted Learning (ARL). In particular, we hypothesize that non-protected features and task labels are valuable for identifying fairness issues, and can be used to co-train an adversarial reweighting approach for improving fairness. Our results show that ARL improves Rawlsian Max-Min fairness, with notable AUC improvements for worst-case protected groups in multiple datasets, outperforming state-of-the-art alternatives.

PDF Details

AAAI Conference 2019 Conference Paper

SNR: Sub-Network Routing for Flexible Parameter Sharing in Multi-Task Learning

Jiaqi Ma
Zhe Zhao
Jilin Chen
Ang Li
Lichan Hong
Ed H. Chi

Machine learning applications, such as object detection and content recommendation, often require training a single model to predict multiple targets at the same time. Multi-task learning through neural networks became popular recently, because it not only helps improve the accuracy of many prediction tasks when they are related, but also saves computation cost by sharing model architectures and low-level representations. The latter is critical for real-time large-scale machine learning systems. However, classic multi-task neural networks may degenerate significantly in accuracy when tasks are less related. Previous works (Misra et al. 2016; Yang and Hospedales 2016; Ma et al. 2018) showed that having more flexible architectures in multi-task models, either manually-tuned or softparameter-sharing structures like gating networks, helps improve the prediction accuracy. However, manual tuning is not scalable, and the previous soft-parameter sharing models are either not flexible enough or computationally expensive. In this work, we propose a novel framework called Sub- Network Routing (SNR) to achieve more flexible parameter sharing while maintaining the computational advantage of the classic multi-task neural-network model. SNR modularizes the shared low-level hidden layers into multiple layers of subnetworks, and controls the connection of sub-networks with learnable latent variables to achieve flexible parameter sharing. We demonstrate the effectiveness of our approach on a large-scale dataset YouTube8M. We show that the proposed method improves the accuracy of multi-task models while maintaining their computation efficiency.

PDF Details

TIST Journal 2015 Journal Article

Who Will Retweet This? Detecting Strangers from Twitter to Retweet Information

Kyumin Lee
Jalal Mahmud
Jilin Chen
Michelle Zhou
Jeffrey Nichols

There has been much effort on studying how social media sites, such as Twitter, help propagate information in different situations, including spreading alerts and SOS messages in an emergency. However, existing work has not addressed how to actively identify and engage the right strangers at the right time on social media to help effectively propagate intended information within a desired time frame. To address this problem, we have developed three models: (1) a feature-based model that leverages people's exhibited social behavior, including the content of their tweets and social interactions, to characterize their willingness and readiness to propagate information on Twitter via the act of retweeting; (2) a wait-time model based on a user's previous retweeting wait times to predict his or her next retweeting time when asked; and (3) a subset selection model that automatically selects a subset of people from a set of available people using probabilities predicted by the feature-based model and maximizes retweeting rate. Based on these three models, we build a recommender system that predicts the likelihood of a stranger to retweet information when asked, within a specific time window, and recommends the top-N qualified strangers to engage with. Our experiments, including live studies in the real world, demonstrate the effectiveness of our work.

Details DOI