Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 20 of 1032 for “"Post-Training"”.
-
Post-Training Large Language Models as Software Engineering Agents
… engineering. This thesis investigates the post-training of LLMs as software engineering agents, focusing on three interconnected challenges: infrastructure, data, and training methodology. First, we contribute to VerlTool, a unified framework for agentic reinforcement learn- ing with tool …
-
Adaptive Post-Training Quantization for Language-Centric Foundation Models
… storage, memory, and inference costs, motivating post-training quantization (PTQ) for efficient deployment. However, architectural and modality-specific differences govern how quantization affects internal representations and downstream performance, motivating complementary quantization approaches …
-
Towards Better Adaptation in Post-Training Large Foundation Models
… objectives nor aligned with human intentions, post-training is required for adapting a raw pretrained model into a helpful and reliable system for practical applications. However, post-training also introduces challenges: (1) Capability Forgetting, where adapting a model for new tasks …
-
PRINCIPLED FRAMEWORKS FOR AI ALIGNMENT: FROM POST-TRAINING TO INFERENCE
… fundamentally different paradigms: The first is post-training alignment, where model parameters are updated after pre-training so that the model better reflects desired behaviors. The second is inference-time alignment, where model behavior is steered at test time without modifying model …
-
Examining the Effects of Post-Training Interventions on Transfer of Training
This study examined the effectiveness of two post-training interventions, goal-setting and self-management training, on improving transfer of training. The interventions provided trainees with cognitive skills and techniques needed to help them transfer their learned skills from the classroom to …
-
Modulation of the Consolidation of Aversive Memory by Post-Training Heroin Administration
… trained on the paradigm, receiving immediate post-training injection of heroin or saline. Conditioned behavior was assessed after 8 training sessions. Clear behavioral differences motivated a control group separation into ‘freezers’ and ‘avoiders’. Heroin treated animals displayed conditioned …
-
Multimodal Learning from Videos: Self-supervised Pre-training, Post-training Alignment, and Benchmarks
… This thesis spans from self-supervised pre-training for learning general and abstract representations, to post-training alignment with subtle human preferences. It further contributes to the design benchmarks to foster progress in multimodal video understanding and reasoning. The first part …
-
An Examination of a Post-Training Strategy to Increase the Transfer of Training
… a problem with the transfer of learning from the training situation to the job situation, suggesting a need to go beyond traditional methods used to increase transfer of training. fhe study presented here extends the research conducted by Erffmeyer (1987) and investigates whether a post-training …
-
A Comprehensive Analysis of the Post-training Effects of Heroin on Memory Consolidation
… thesis aimed to test the hypothesis that the post-training effects of heroin are mediated by the type of memory that a task relies on. To test this, male Sprague-Dawley rats undergoing heroin or cocaine place conditioning received post-training injections of 1 mg/kg heroin. It was found that …
-
Symmetric Post-Training Static Quantization for Radial Basis Function-based Kolmogorov-Arnold Networks
Τα Δίκτυα Kolmogorov-Arnold (KAN) εμφανίστηκαν πρόσφατα ως εναλλακτική στα παραδοσιακά multi-layer perceptron (MLP) μοντέλα, αντικαθιστώντας τις σταθερές συναρτήσεις ενεργοποίησης με εκπαιδεύσιμες. Η ανάπτυξή τους σε πλατφόρμες με περιορισμένους πόρους, ωστόσο, απαιτεί αποτελεσματική συμπίεση. Η …
-
Reliable and efficient post-training of large reasoning models : reinforcement learning dynamics and adaptation
… models are increasingly improved through post-training rather than pre-training alone, but strong post-training requires more than simply maximizing task reward. This dissertation studies reliable and efficient post-training of large reasoning models, with a focus on reinforcement learning …
-
Sustaining the effect of gatekeeper training: an empirical examination of two post-training interventions
… supporting the application of gatekeeper training programs toward suicide prevention efforts (e.g., Cross, Mattieu, Cerel, & Knox, 2007; Wyman, Inman, Brown, Cross, Schmeelk-Cone, & Pena, 2008). However, recent studies (e.g., Chen, Gibbs, & Moore, 2009; Keller et al., 2009) suggest that …
-
Analysis of the Use of Internet -Based Communication Technology for Online Post -Training Coaching
Several implications for trainers and training managers, technology coordinators, and researchers emerged. Training professionals should select communication media based on richness and message equivocality and should develop activities that promote individual and group communication. Technology …
-
Post-Training Optimization of Cross-layer Approximate Computing for Edge Inference of Deep Learning Applications
… unit (ACU) simulation for evaluation and training of approximated DNNs. ProxSim supports quantization and approximation of common neural layers such as fully connected (FC), convolutional, and recurrent layers. A performance evaluation using a variety of DNN architectures, as well as a …
-
Deep Reinforcement Learning for Data-Agnostic Post-Training Debiasing of Black-Box Machine Learning Models
… the models’ internal architectures, the models’ training datasets, or both. In this paper we outline the reasons why such requirements are disadvantageous, and present an alternative novel debiasing system that is both data- and model-agnostic. We implement this system as a Reinforcement Learning …
-
The impact of individual differences on the interviewing success and post-training performance of investigative interviewers
… investigative interviewing and the efficacy of training was examined in a three-step research design. In Study 1, the structure of a 50-item aptitudes scale, a modified and extended version of the Police Interviewing Competences Inventory (PICI), was assessed using a general population sample (N …
-
POST-TRAINING OF VISION-LANGUAGE AGENTS FOR DECENTRALIZED AUTONOMOUS VEHICLE COORDINATION USING GENERALIZABLE MULTI-AGENT REWARDS
… addresses this gap by introducing a social post-training method for Alpamayo-R1 (AR1) that explicitly rewards behavior that is predictable to neighboring agents. We extend AR1’s Group Relative Policy Optimization (GRPO) post-training by augmenting thereward with Expectation Alignment …
Page 1 of 52