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 9 of 9 for “"Test-time scaling"”.
-
ADVANCING TEST-TIME COMPUTE SCALING IN LANGUAGE MODELS: FROM TRAINING TO INFERENCE
Large language models increasingly benefit from scaling test-time computation, but effective scaling requires more than simply allocating additional inference compute. This thesis studies four complementary bottlenecks across the model lifecycle: context capacity, reasoning capability, efficient …
-
PRINCIPLED FRAMEWORKS FOR AI ALIGNMENT: FROM POST-TRAINING TO INFERENCE
… desired behaviors. The second is inference-time alignment, where model behavior is steered at test time without modifying model parameters. While both paradigms aim to improve alignment, they present distinct challenges and opportunities.This thesis advances AI alignment across both of these …
-
A Comparative Study of Inference-Time Scaling Strategies for Large Language Models
… size or training data, but also on inference-time strategies. However, existing inference-time methods are typically evaluated in isolation and under differing experimental assumptions, making it difficult to draw systematic conclusions about their relative effectiveness. This thesis proposes …
-
INVESTIGATING REASONING MODELS: OVERTHINKING AND THINKING EPISODE THEORY
… We show that this phenomenon contradicts the test-time scaling law: more computation does not help models identify the missing information. Surprisingly, non-reasoning models demonstrate better critical thinking ability in these scenarios. Second, we introduce a novel analytical framework …
-
Code Generation with Large Language Models: Inductive Reasoning and Calibration
… data. Furthermore, we demonstrate how scaling test-time computation significantly improves inference results in this PBE setting. Our approach achieves state-of-the-art results on common PBE benchmarks including string, number sequence, and logo graphics domains. We further extend our …
-
Optimization in the Era of Large Language Models: Theory and Applications
… data curation, class-based error analysis, and test-time scaling, significantly boosting formulation robustness on realistic benchmarks. In Part II (Chapters 6-7), we shift our focus to the rigorous theoretical analysis of optimization algorithms. Chapter 6 investigates the smoothed complexity …
-
Artificial General Intelligence (AGI)-Native Wireless Systems: Digital Twins and World Models for Beyond 6G Networks
… that ensures the optimal average synchronization time of the world models, while satisfying the synchronization intensity requirements of the DTs. Results show that synchronization delays can be reduced up to 25 % in comparison to the standard signal-to noise ratio (SNR) association benchmark. …
-
STEERING GENERATIVE AI ON THE FLY: INFERENCE-TIME APPROACHES FOR SAFE, RELIABLE, AND INCLUSIVE LANGUAGE MODELS
… aligned with human values through training-time methods such as reinforcement learning from human feedback. However, these methods produce static policies that cannot adapt to adversarial inputs unseen during training, reasoning challenges that emerge at test time, or the linguistic …
-
Co-architecting scalable intelligence: from hardware substrates to algorithms
… runs and increasingly compute-intensive test-time scaling. However, despite these innovations, the prevailing “brute-force” trajectory—relying on ever larger models and ever deeper compute budgets—is becoming unsustainable. In the post-Moore era, a widening gap is emerging between rapidly …