Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks147 from public data
- MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents231
MEM1, an end-to-end reinforcement learning framework that enables agents to operate with constant memory across long multi-turn tasks, is introduced, an end-to-end reinforcement learning framework that enables agents to operate with constant memory across long multi-turn tasks.
- Goat: Fine-tuned LLaMA Outperforms GPT-4 on Arithmetic Tasks109
Goat, a fine-tuned LLaMA model that significantly outperforms GPT-4 on a range of arithmetic tasks, is introduced and an approach that classifies tasks based on their learnability, and subsequently decomposes unlearnable tasks into a series of learnable tasks by leveraging basic arithmetic principles is proposed.
- A Unifying Framework of Anytime Sparse Gaussian Process Regression Models with Stochastic Variational Inference for Big Data95
This paper presents a novel unifying framework of anytime sparse Gaussian process regression (SGPR) models that can produce good predictive performance fast and improve their predictive performance over time and empirically evaluates the trade-off between the predictive performance vs. time efficiency of the anytime SGPR models on two real-world million-sized datasets.
- ReasonIR: Training Retrievers for Reasoning Tasks88
ReasonIR-8B is presented, the first retriever specifically trained for general reasoning tasks, and its performance consistently increases with longer and more information-rich rewritten queries; it continues to outperform other retrievers when combined with an LLM reranker.
- Data Valuation in Machine Learning: "Ingredients", Strategies, and Open Challenges87
A comprehensive technical survey is presented that provides a new formal study of data valuation in ML through its “ingredients” and the corresponding properties, and grounds the discussion of common desiderata satisfied by existing data valuation strategies on their proposed ingredients.
- FCM-sketch85
This work proposes FCM, a framework that is designed to support generic network measurement with high accuracy and can reduce the errors in many measurement tasks by 50% to 80% compared to CM-Sketch and other state-of-the-art approaches.
- Bayesian Optimization Meets Bayesian Optimal Stopping71
This paper proposes to unify BO (specifically, Gaussian process-upper confi-dence bound (GP-UCB) with Bayesian optimal stopping (BO-BOS) with Bayesian optimal stopping (BO-BOS) to boost the epoch efficiency of BO.
- A distributed variational inference framework for unifying parallel sparse Gaussian process regression models60
A novel distributed variational inference framework that unifies many parallel sparse Gaussian process regression (SGPR) models for scalable hyperparameter learning with big data and empirically evaluates the predictive performance and scalability of the distributedvariational SGPR models unified by this framework on two real-world datasets.
- PINNACLE: PINN Adaptive ColLocation and Experimental points selection58
This work theoretically shows that the criterion used by PINNACLE is related to the PINN generalization error, and empirically demonstrate that PINNACLE is able to outperform existing point selection methods for forward, inverse, and transfer learning problems.
- 51
- Distributed Batch Gaussian Process Optimization50
Empirical evaluation on synthetic benchmark objective functions and a real-world optimization problem shows that DB-GP-UCB outperforms the state-of-the-art batch BO algorithms.
- CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery49
Evaluated on diverse mathematical, algorithmic, and systems optimization tasks, CORAL sets new state-of-the-art results on 10 tasks, achieving 3-10 times higher improvement rates with far fewer evaluations than fixed evolutionary search baselines across tasks.
- Markov Chain Monte Carlo-Based Machine Unlearning49
A Markov chain Monte Carlo-based machine unlearning (MCU) algorithm that helps to effectively and efficiently unlearn a trained model from subsets of training dataset and can be used to erase the lineage of a user's personal data from trained ML models, thus upholding a users' "right to be forgotten".
- Prompt Optimization with Human Feedback47
The problem of prompt optimization with human feedback (POHF), in which the APOHF aims to optimize the prompt for a black-box LLM using only human preference feedback, is studied.
- Incentivizing Collaboration in Machine Learning via Synthetic Data Rewards44
This paper proposes a data valuation function using maximum mean discrepancy (MMD) that values data based on its quantity and quality in terms of its closeness to the true data distribution and formulate the reward scheme as a linear optimization problem that guarantees certain incentives such as fairness in the CGM framework.
- 44
- Towards Robust ResNet: A Small Step but a Giant Leap44
It is proved that a small step factor h can benefit the training robustness for back-propagation and the generalization robustness of the residual network by an explicit Euler method.
- Unifying and Boosting Gradient-Based Training-Free Neural Architecture Search41
A unified theoretical analysis of gradient-based training-free NAS is presented, which allows us to theoretically study their relationships, theoretically guarantee their generalization performances, and exploit the unified theoretical understanding to develop a novel framework named hybrid NAS (HNAS), which consistently boosts training- free NAS in a principled way.
- Collective Model Fusion for Multiple Black-Box Experts35
This paper presents the first collective model fusion framework for multiple experts with heterogeneous black-box architectures, addressing the key issues of how black- box experts interact to understand the predictive behaviors of one another and how the shared understandings can be combined to generate high-quality consensus prediction.
- Top-k Ranking Bayesian Optimization31
This paper designs a surrogate model that is not only capable of catering to the above observations, but is also supported by a classic random utility model and introduces the first information-theoretic acquisition function in BO with preferential observation called multinomial predictive entropy search (MPES).
- 31
- 31
- Localized Zeroth-Order Prompt Optimization30
A novel algorithm, namely localized zeroth-order prompt optimization (ZOPO), which incorporates a Neural Tangent Kernel-based derived Gaussian process into standard zeroth-order optimization for an efficient search of well-performing local optima in prompt optimization is proposed.
- Use Your INSTINCT: INSTruction optimization for LLMs usIng Neural bandits Coupled with Transformers28
A neural bandit algorithm is adopted which replaces the GP in BO by an NN surrogate to optimize instructions for black-box LLMs and is used to propose the INSTINCT algorithm, which consistently outperforms the existing methods in different tasks.
- Gaussian process decentralized data fusion meets transfer learning in large-scale distributed cooperative perception26
A novel transfer learning mechanism for a team of agents capable of sharing and transferring information encapsulated in a summary based on a support set to that utilizing a different support set with some loss that can be theoretically bounded and analyzed is proposed.
- Waterfall: Scalable Framework for Robust Text Watermarking and Provenance for LLMs25
It is empirically demonstrate that Waterfall achieves significantly better scalability, robust verifiability, and computational efficiency compared to SOTA article-text watermarking methods, and also showed how it could be directly applied to the watermarking of code.
- An Information-Theoretic Framework for Unifying Active Learning Problems25
A novel active learning criterion is introduced that subsumes an existing LSE algorithm and achieves state-of-the-art performance in LSE problems with a continuous input domain and a competitive information-theoretic acquisition function for BO that has interesting connections to upper confidence bound and max-value entropy search.
- Learning Task-Agnostic Embedding of Multiple Black-Box Experts for Multi-Task Model Fusion25
A new fusion paradigm is developed that represents each expert as a distribution over a spectrum of predictive prototypes, which are isolated from task-specific information encoded within the prototype distribution and can then be reintegrated to generate a new model that solves a new task encoded with a different prototype distribution.
- Scalable Variational Bayesian Kernel Selection for Sparse Gaussian Process Regression24
Though it is computationally challenging to jointly optimize a large number of hyperparameters due to many kernels being evaluated simultaneously by the VBKS algorithm, it is shown that the variational lower bound of the log-marginal likelihood can be decomposed into an additive form such that each additive term depends only on a disjoint subset of the Variational variables and can thus be optimized independently.
- Stochastic Variational Inference for Bayesian Sparse Gaussian Process Regression24
A novel variational inference framework for deriving a family of Bayesian sparse Gaussian process regression models whose approximations are variationally optimal with respect to the full-rank GPR model enriched with various corresponding correlation structures of the observation noises.
- Artificial Intelligence Research in Singapore: Assisting the Development of a Smart Nation24
Artificial Intelligence (AI) research in Singapore is focused on accelerating the country’s development into a Smart Nation and in developing automated methods and systems to improve quality of life.
- Source Attribution for Large Language Model-Generated Data21
This paper shows that both problems can be solved by watermarking, i.e., by enabling an LLM to generate synthetic texts with embedded watermarks that contain information about their source(s) and achieves effective source attribution and data provenance.
- Waterfall: Framework for Robust and Scalable Text Watermarking and Provenance for LLMs20
It is demonstrated that Waterfall achieves significantly better scalability, robust verifiability, and computational efficiency compared to SOTA article-text watermarking methods, and showed how it could be directly applied to the watermarking of code.
- Bayesian Optimization under Stochastic Delayed Feedback20
This paper proposes algorithms with sub-linear regret guarantees that efficiently address the dilemma of selecting new function queries while waiting for randomly delayed feedback.
- Helpful or Harmful Data? Fine-tuning-free Shapley Attribution for Explaining Language Model Predictions19
This work introduces an efficient fine-tuning-free approximation of the Shapley value (FreeShap) for instance attribution based on the neural tangent kernel and empirically demonstrates that FreeShap outperforms other methods for instance attribution and other data-centric applications such as data removal, data selection, and wrong label detection.
- 19
- Fair yet Asymptotically Equal Collaborative Learning17
This paper explores an incentive design that guarantees fairness so that nodes receive rewards commensurate to their contributions and empirically demonstrates that the proposed approach outperforms existing baselines in fairness and learning performance while remaining competitive in preserving equality.
- Trade-Off Between Payoff and Model Rewards in Shapley-Fair Collaborative Machine Learning15
This paper proposes desirable properties for achieving a fair adjustment of the payoff flows that can trade off between the model reward’s performance and the payoff reward.
- On the Convergence of the Shapley Value in Parametric Bayesian Learning Games15
This paper establishes the convergence property of the Shapley value in parametric Bayesian learning games where players perform a Bayesian inference using their combined data, and the posterior-prior KL divergence is used as the characteristic function.
- Federated Neural Bandits14
The federated neural-upper confidence bound (FN-UCB) algorithm is introduced, which adopts a weighted combination of two UCBs: UCB a allows every agent to additionally use the observations from the other agents to accelerate exploration (without sharing raw observations); UCB b uses an NN with aggregated parameters for reward prediction in a similar way as federated averaging for supervised learning.
- On Provably Robust Meta-Bayesian Optimization13
It is proved that both meta-BO algorithms are asymptotically no-regret even when some or all previous tasks are dissimilar to the current task, and it is shown that RM-GP-UCB enjoys a better theoretical robustness than RM- GP-TS.
- AID: Active Distillation Machine to Leverage Pre-Trained Black-Box Models in Private Data Settings13
This paper presents an active distillation method for a local institution to find the best queries within its given budget to distill an on-server black-box model’s predictive knowledge into a local surrogate with transparent parameterization, which addresses several challenges of deploying machine learning in many industrial settings with strong proprietary constraints.
- 13
- ActiveDPO: Active Direct Preference Optimization for Sample-Efficient Alignment12
This work proposes an algorithm, ActiveDPO, that uses a theoretically grounded data selection criterion for non-linear reward functions while directly leveraging the LLM itself to parameterize the reward model used for active data selection, thereby leading to more effective and efficient data collection.
- Understanding Domain Generalization: A Noise Robustness Perspective12
Finite-sample analysis reveals that label noise exacerbates the effect of spurious correlations for ERM, undermining generalization, and illustrates that DG algorithms exhibit implicit label-noise robustness during finite-sample training even when spurious correlation is present.
- Adjusted Expected Improvement for Cumulative Regret Minimization in Noisy Bayesian Optimization12
A high-probability regret upper bound of EIC is established based on the maximum information gain which is tighter than the bound of existing EI-based algorithms and comparable to the regret bound of other popular BO algorithms such as Thompson sampling and upper confidence bound.
- Robustifying and Boosting Training-Free Neural Architecture Search11
The robustifying and boosting training-free NAS (RoBoT) algorithm is proposed which employs the optimized combination of existing training-free metrics explored from Bayesian optimization to develop a robust and consistently better-performing metric on diverse tasks, and applies greedy search on the newly developed metric to bridge the gap.
- Ferret: Federated Full-Parameter Tuning at Scale for Large Language Models11
Ferret is proposed, the first first-order method with shared randomness to enable scalable full-parameter tuning of LLMs across decentralized data sources while maintaining competitive model accuracy and high computational efficiency, reduced communication overhead, and fast convergence.
- Federated Zeroth-Order Optimization using Trajectory-Informed Surrogate Gradients11
The FZooS achieves theoretical improvements over the existing approaches, which is supported by real-world experiments such as federated black-box adversarial attack and federated non-differentiable metric optimization.
- REFRAG: Rethinking RAG based Decoding9
It is argued that most computations over the RAG context during decoding are unnecessary and can be eliminated with minimal impact on performance, and proposed REFRAG, an efficient decoding framework that compresses, senses, and expands to improve latency in RAG applications.
- WaterDrum: Watermarking for Data-centric Unlearning Metric9
This paper presents the first data-centric unlearning metric for LLMs called WaterDrum that exploits robust text watermarking for overcoming limitations and introduces new benchmark datasets for LLM unlearning that contain varying levels of similar data points and can be used to rigorously evaluate unlearning algorithms using WaterDrum.
- DETAIL: Task DEmonsTration Attribution for Interpretable In-context Learning9
This work proposes an influence function-based attribution technique, DETAIL, that addresses the specific characteristics of ICL and shows how DETAIL can help improve model performance in real-world scenarios through demonstration reordering and curation.
- Bayesian Optimization with Cost-varying Variable Subsets9
This paper presents a novel Gaussian process upper confidence bound-based algorithm for solving the BOCVS problem that is provably no-regret and empirically shows that the proposed algorithm can find significantly better solutions than comparable baselines with the same budget.
- 9
- Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volume8
UMPIRE is introduced, a training-free uncertainty quantification framework for MLLMs that works efficiently across various input and output modalities without external tools, relying only on the models' own internal modality features.
- DUET: Optimizing Training Data Mixtures via Feedback from Unseen Evaluation Tasks8
DUET is presented, a novel global-to-local algorithm that interleaves influence function as a data selection method with Bayesian optimization to optimize data mixture via feedback from a specific unseen evaluation task to optimize training data mixtures via feedback from an unseen evaluation task.
- Data Distribution Valuation8
This work proposes a maximum mean discrepancy (MMD)-based valuation method which enables theoretically principled and actionable policies for comparing data distributions from samples, and empirically demonstrates that the method is sample-efficient and effective in identifying valuable data distributions against several existing baselines.
- 8
- Model Shapley: Equitable Model Valuation with Black-box Access8
This work investigates the black-box access setting which allows querying a model (to observe predictions) without disclosing model-specific information (e.g., architecture and parameters) and proposes a novel and equitable model valuation method called model Shapley.
- Training-Free Neural Active Learning with Initialization-Robustness Guarantees8
This work empirically demonstrate that the expected variance with Gaussian processes (EV-GP) criterion is highly correlated with both initialization robustness and generalization performance, and show that it consistently outperforms baseline methods in terms of both desiderata, especially in situations with limited initial data or large batch sizes.
- Recursive reasoning-based training-time adversarial machine learning8
It is shown how an R2T2 attacker (defender) can utilize the proposed nested projected gradient descent-based method to approximate the optimal attack (defense) strategy at an arbitrary level of reasoning.
- 8
- 7
- Pruning during training by network efficacy modeling7
Empirical evaluations show that the proposed Bayesian early pruning improves the computational efficiency of DNN training while better preserving model performance compared to other tested pruning approaches.
- 7
- 7
- On Newton's Method to Unlearn Neural Networks6
CureNewton's method is proposed, a principle approach that leverages cubic regularization to handle the Hessian degeneracy effectively and can achieve competitive unlearning performance to the state-of-the-art algorithm in practical unlearning settings, while being theoretically justified and efficient in running time.
- Convolutional Normalizing Flows for Deep Gaussian Processes6
A novel convolutional normalizing flow (CNF) is developed to improve the time efficiency and capture dependency between layers and empirical evaluation demonstrates that CNF DGP outperforms the state-of-the-art approximation methods for DGPs.
- TreeGrad-Ranker: Feature Ranking via $O(L)$-Time Gradients for Decision Trees5
The use of probabilistic values, which include the well-known Shapley and Banzhaf values, is revisited and TreeGrad-Ranker, which aggregates the gradients while optimizing the joint objective to produce feature rankings, and TreeGrad-Shap, a numerically stable algorithm for computing Beta Shapley values with integral parameters are introduced.
- Gradient-Free Methods for Nonconvex Nonsmooth Stochastic Compositional Optimization5
This paper proposes gradient-free stochastic methods for finding the ( δ, ϵ ) -Goldstein stationary points of such problems with non-asymptotic convergence rates and leads to an improved convergence rate for the convex nonsmooth SCO problem.
- Data-Centric AI in the Age of Large Language Models5
A data-centric viewpoint of AI research, focusing on large language models (LLMs), is proposed, covering data-centric benchmarks and data curation, data attribution, knowledge transfer, knowledge transfer, and inference contextualization.
- MeMo: Memory as a Model4
MeMo (Memory as a Model), a modular framework that encodes new knowledge into a dedicated memory model while keeping the LLM parameters unchanged, is introduced, showing that MeMo achieves strong performance compared to existing methods across diverse settings.
- BILBO: BILevel Bayesian Optimization4
BILevel Bayesian Optimization (BILBO), a novel Bayesian optimization algorithm for general bilevel problems with blackbox functions, which optimizes both upper- and lower-level problems simultaneously, without the repeated lower-level optimization required by existing methods is presented.
- Active Human Feedback Collection via Neural Contextual Dueling Bandits4
It is theoretically show that when preference feedback follows the Bradley-Terry-Luce model, the worst sub-optimality gap of the policy learned by Neural-ADB decreases at a sub-linear rate as the preference dataset increases.
- TRACE: TRansformer-based Attribution using Contrastive Embeddings in LLMs4
This work proposes a novel and versatile TRansformer-based Attribution framework using Contrastive Embeddings called TRACE that exploits contrastive learning for source attribution and shows that TRACE significantly improves the ability to attribute sources accurately, making it a valuable tool for enhancing the reliability and trustworthiness of LLMs.
- 4
- MineDraft: A Framework for Batch Parallel Speculative Decoding3
MineDraft, a batch parallel speculative decoding framework designed to effectively hide drafting latency by overlapping it with verification, is proposed, and the theoretical analysis shows that PSD is substantially more efficient than standard SD.
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- Is Data Shapley Not Better than Random in Data Selection? Ask NASH1
A novel data selection framework, NASH (Non-linear Aggregation of SHapley-informative components), which decomposes the target utility function into simpler, Shapley-informative component functions, and selects data by optimizing an objective that aggregates these components non-linearly.
- BoLT: A Benchmark to Democratize Black-box Optimization Research for Expensive LLM Tasks1
BoLT is introduced, the first LLM-centric benchmark that democratizes LLM research for the BBO community, and benchmark BoLT against an extensive range of BO and BBO methods, showing that selected BO methods consistently outperform others across tasks and highlighting gaps in existing BBO methods on LLM tasks.
- PC-MoE: memory-efficient and privacy-preserving collaborative training for Mixture-of-Experts LLMs1
This work introduces Privacy-preserving Collaborative Mixture-of-Experts (PC-MoE), which leverages the sparsity of the MoE architecture for memory-efficient decentralized collaborative LLM training, enabling multiple parties with limited GPU-memory and data resources to collectively train more capable LLMs than they could achieve individually.
- Weight-Informed Self-Explaining Clustering for Mixed-Type Tabular Data1
WISE is proposed, a Weight-Informed Self-Explaining framework that unifies representation, feature weighting, clustering, and interpretation in a fully unsupervised and transparent pipeline and produces faithful, human-interpretable explanations grounded in the same primitives that drive clustering.
- 1
- De-attribute to Forget for LLM Unlearning1
This paper proposes the first LLM unlearning framework based on data attribution rewards called DareU that performs reinforcement learning to update the LLM by reducing the attribution score of its generated responses (i.e., de-attributing) to the forget data owners.
- Broaden your SCOPE! Efficient Multi-turn Conversation Planning for LLMs with Semantic Space1
A novel approach called Semantic space COnversation Planning with improved Efficiency (SCOPE) that exploits the dense semantic representation of conversations to perform conversation planning efficiently and can perform conversation planning 70 times faster than conventional simulation-based planning algorithms.
- 1
- COBRA: Contextual Bandit Algorithm for Ensuring Truthful Strategic Agents1
This paper proposes an algorithm, COBRA, for contextual bandit problems involving strategic agents that disincentivize their strategic behavior without using any monetary incentives, while having incentive compatibility and a sub-linear regret guarantee.
- Active Set Ordering1
This paper formalizes the active set ordering problem, which involves actively discovering a set of inputs based on their orderings determined by expensive evaluations of a blackbox function, and proposes the mean prediction algorithm, which is theoretically analyzed in terms of the regret of predicted pairwise orderings between inputs.
- 1
- 1
- Data value estimation on private gradients1
This work proposes to instead inject carefully correlated noise to provably remove the linear scaling of estimation uncertainty w.r.t.~the budget and shows that this method gives better data value estimates on various ML tasks and is applicable to use cases including dataset valuation and FL.
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- Diffusion-Inspired Reconfiguration of Transformers for Uncertainty Calibration–
A diffusion-inspired reconfiguration of transformers in which each feature transformation block is modeled as a probabilistic mapping that mimics the structure of a diffusion process, transporting data mass from the input distribution to the pre-trained feature distribution.
- Nonlinear Axiomatic Attribution for Cooperative Games–
Inspired by the least core, which is a popular nonlinear substitute for the Shapley value, a class of nonlinear attribution methods that retain the remaining necessary axioms is introduced that aims to approximate utility functions as faithfully as possible.
- The Chicken and Egg Dilemma: Co-optimizing Data and Model Configurations for LLMs–
JoBS is introduced, an approach that uses a scaling-law-inspired performance predictor to aid Bayesian optimization (BO) in jointly optimizing LLM training data and model configurations efficiently and outperforms existing multi-fidelity BO baselines, as well as data and model optimization approaches across diverse LLM tasks under the same optimization budget.
- –
- –
- –
- Incentivizing Truthfulness and Collaborative Fairness in Bayesian Learning–
This paper presents the first mechanism that provably ensures collaborative fairness and incentivizes data truthfulness at equilibrium for Bayesian models and discusses the implications and suitable relaxations when the mediator has a limited budget for rewards or lacks a validation set.
- Adalina: Adaptive Linear Approximation for the Shapley Value and Beyond–
This work introduces the first adaptive, linear-time, linear-space randomized algorithm, Adalina, that theoretically achieves improved mean square error and establishes a theoretical framework that enables sharper query complexities for existing unbiased randomized algorithms.
- –
- –
- INO-SGD: Addressing Utility Imbalance under Individualized Differential Privacy–
The INO-SGD algorithm is proposed, which strategically down-weights data within each batch to improve performance on the more private data across all iterations, and is specially designed to satisfy IDP.
- –
- –
- BarrierSteer: LLM Safety via Learning Barrier Steering–
This paper introduces BarrierSteer, a novel inference-time framework that improves response safety by embedding learned nonlinear safety constraints directly into the model's latent representation space, enabling constraint-guided steering of unsafe latent trajectories during generation.
- Incentivizing Black-Box Model Sharing with Fair Rewards and Payoffs–
This work proposes a novel incentive mechanism that fairly distributes ensemble predictions and monetary payoffs commensurate with each agent's contribution and payment and uses the average ensemble weight for the contribution measure.
- ExTra: Exploratory Trajectory Optimization for Language Model Reinforcement Learning–
This work introduces ExTra (Exploratory Trajectory Optimization), a GRPO-compatible framework that extracts exploration signals from the model's own rollouts and combines two mechanisms: a novelty reward that adds embedding-based diversity bonuses after GRPO normalization, rewarding diverse correct solutions.
- How Hard Can It Be? Hardness-Aware Multi-Objective Unlearning–
An unlearning algorithm (HAMU) is derived with the overall goal of guaranteeing a specified improvement in forget quality while minimizing the retain utility cost/degradation by updating the model weights based on the authors' hardness measure.
- –
- –
- –
- –
- –
- –
- Self-Interested Agents in Collaborative Machine Learning: An Incentivized Adaptive Data-Centric Framework–
This framework is underpinned by non-asymptotic analyses, ensuring convergence of the agent-side policy optimization to an approximate stationary point of the evaluation functions and convergence of the arbiter-side optimization to an approximate stationary point of the expected loss function.
- –
- –
- –
- Global-to-Local Support Spectrums for Language Model Explainability–
This paper proposes a method to generate an explanation in the form of support spectrums which is able to generate explanations that are tailored to specific test points and shows the effectiveness of the method in image classification and text generation tasks.
- –
- –
- –
- –
- –
- –
- Trusted Data Sharing - Incentivizing Collaboration and Rights to be Forgotten in Machine Learning–
This talk will discuss how to perform data valuation, incentivize multiple parties with data to collaborate in building higher-quality models, and unlearn a trained machine learning model from data to be erased for compliance with regulations such as the Personal Data Protection Act, General Data Protection Regulation, and the future of personal data ownership.
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-10. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it. You will be asked to sign in.