uk

Enoch H. Kang

Best AI papers explained

Cut through the noise. We curate and break down the most important AI papers so you don’t have to.

Обов'язково відвідайте сайт подкасту та підтримайте його автора: podcasters.spotify.com

Автор

Enoch H. Kang

Категорія

Technology

Сайт подкасту

podcasters.spotify.com

Останній епізод

19 лис 2026

Де слухати?

Подкасти в застосунку Replaio Radio Уже незабаром

Подкасти незабаром з'являться в застосунку. Встановіть уже зараз і першими побачте зовсім новий погляд на подкасти

Завантажити з Google Play Встановіть безкоштовно Android 10 млн+ завантажень · рейтинг 4,7+ iOS незабаром

Епізоди

Matching Features, Not Tokens: Energy-Based Fine-Tuning of Language Models 16.03.2026

This research paper introduces Energy-Based Fine-Tuning (EBFT), a novel method for refining language models by matching feature statistics of generated text with ground-truth data. Traditional training relies on next-token prediction, which often causes models to drift or fail during long sequences because they lack global distributional calibration. By optimizing a feature-matching objective usin...

Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights 14.03.2026

This research introduces the concept of Neural Thickets, describing a phenomenon where large pretrained models are surrounded by a high density of diverse, task-specific solutions in their local weight space. While small models require structured optimization like gradient descent to find improvements, larger models transition into a regime where random weight perturbations frequently yield "...

AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization 14.03.2026

This paper introduces AdaEvolve, a novel framework designed to enhance how Large Language Models (LLMs) solve complex optimization and programming tasks through evolutionary search. Unlike existing methods that use rigid, pre-set schedules, this system implements hierarchical adaptivity to manage computational resources and search strategies dynamically. It operates across three levels: local adap...

∇−reasoner: LLM reasoning via test-time gradient descent in latent space 14.03.2026

This paper introduces ∇-Reasoner, a novel framework that improves Large Language Model (LLM) reasoning by applying gradient-based optimization during the inference process. Unlike traditional methods that rely on random sampling or discrete searches, this approach uses Differentiable Textual Optimization (DTO) to refine token logits through first-order gradients derived from reward models and like...

Inference for Regression with Variables Generated by AI or Machine Learning 12.03.2026

This research investigates how using artificial intelligence (AI) or machine learning (ML) to generate variables for economic regressions can lead to biased estimates and invalid statistical inference. While researchers often treat AI-generated outputs as standard data, the authors demonstrate that measurement error in these variables—even from high-performance algorithms—shifts the centering of c...

Fast KV Compaction via Attention Matching 12.03.2026

This paper introduces Attention Matching (AM), a novel framework for fast and efficient key-value (KV) cache compaction in long-context language models. As models process longer sequences, the memory required for the KV cache becomes a major bottleneck, often necessitating lossy strategies like summarization or token eviction. The researchers propose optimizing compact keys and values to reproduce...

Position: stop anthropomorphizing intermediate tokens as reasoning/thinking traces! 11.03.2026

This position paper argues against the anthropomorphization of intermediate tokens in large language models, commonly referred to as "reasoning traces" or "chains of thought." The authors contend that these outputs are not genuine reflections of human-like thinking but are instead statistically generated patterns that may lack semantic validity. Research indicates that model pe...

Code World Models for General Game Playing 08.03.2026

Researchers at Google DeepMind introduced Code World Models (CWM), a framework that uses Large Language Models to translate natural language game rules and player trajectories into executable Python code. Unlike traditional methods that use LLMs as direct move-generating policies, this approach treats the model as a verifiable simulation engine capable of defining state transitions and legal actio...

Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought 07.03.2026

This research paper explores how Chain of Thought (CoT) prompting enables transformers to solve complex mathematical problems by mimicking iterative optimization techniques. The authors demonstrate that while standard models are limited to a single stage of calculation, using intermediate reasoning steps allows a transformer to execute multi-step gradient descent internally. Through the lens of li...

Task Descriptors Help Transformers Learn Linear Models In-Context 07.03.2026

This paper explores how task descriptors, such as a mean value $\mu$, improve in-context learning (ICL) for linear regression within Transformer models. By examining a one-layer linear self-attention (LSA) network, the researchers demonstrate that models can effectively utilize these descriptors to standardize input data and reduce prediction errors. The paper provides a mathematical proof that gr...

Equivalence of Context and Parameter Updates in Modern Transformer Blocks 07.03.2026

This research explores how modern Large Language Models adapt to new information during inference by framing in-context learning as a series of implicit weight updates. The authors demonstrate that the influence of a prompt can be mathematically mapped to specific, rank-1 patches on a model's existing parameters, effectively "reprogramming" the network without formal retraining. By e...

Learning without training: The implicit dynamics of in-context learning 07.03.2026

This research explores the mechanisms of in-context learning (ICL) in Large Language Models, proposing that transformers learn by implicitly updating their internal weights during inference. The authors demonstrate that a transformer block effectively transforms prompt examples into a rank-1 weight update of the model's MLP layer. This process allows the model to adapt to new patterns without...

Causal Identification from Counterfactual Data: Completeness and Bounding Results 07.03.2026

This research establishes a complete algorithmic framework for identifying counterfactual quantities by utilizing a newly discovered family of physically realizable Layer 3 data. While traditional causal inference was restricted to observational and interventional data, the authors introduce the CTFIDU+ algorithm, which can determine if a counterfactual query is identifiable from arbitrary sets of...

Is Cosine-Similarity of Embeddings Really About Similarity? 06.03.2026

This paper investigates whether cosine similarity accurately reflects the semantic similarity of learned embeddings, particularly in linear matrix factorization models. The authors demonstrate that the metric can produce arbitrary or non-unique results because certain training objectives allow for the random rescaling of latent dimensions. While some regularization methods yield a unique solution,...

Diffusion LLMs are Natural Adversaries for any LLM 05.03.2026

This research introduces **INPAINTING**, a framework that treats finding adversarial "jailbreak" prompts as a simple inference task rather than a slow optimization problem. By using **Diffusion Large Language Models (DLLMs)**, which understand the joint relationship between prompts and responses, the researchers can directly generate prompts that trigger specific harmful outputs. This method effec...

Are you going to finish that? A Practical Study of the Partial Token Problem 04.03.2026

Language models face a significant partial token problem (PTP) when user prompts end in the middle of a multi-character token, causing the model to misinterpret the expected continuation. This research highlights that the issue is not just a theoretical glitch but a pervasive failure mode in natural language use, especially in Chinese, German, and programming code where word and token boundaries f...

Language Models Struggle to Use Representations Learned In-Context 02.03.2026

This paper investigates whether large language models (LLMs) can effectively utilize novel information they learn during a single conversation. While models successfully create internal representations of new concepts from context, the study reveals these representations are often inert and cannot be applied to downstream tasks. Experiments on next-token prediction and adaptive world modeling show...

LLMs are Bayesian, In Expectation, Not in Realization 01.03.2026

This research explores the discrepancy between transformer in-context learning and Bayesian inference, arguing that models are Bayesian in expectation rather than through every individual realization. While previous studies used martingale diagnostics to question the Bayesian nature of these models, this paper identifies positional encodings as the primary factor that breaks the required exchangea...

Learning from Trials and Errors: Reflective Test-Time Planning for Embodied LLMs 27.02.2026

This paper discusses a framework for reflective test-time planning designed to improve the performance of embodied Large Language Models (LLMs) during robotic tasks. This system utilizes double-loop learning, where agents re-evaluate their past decisions through hindsight assessments to correct underlying strategic errors. By incorporating internal reflection for immediate scoring and retrospectiv...

LLMs Can Learn to Reason Via Off-Policy RL 27.02.2026

This research introduces Optimal Advantage-based Policy Optimization with Lagged Inference policy (OAPL), a novel reinforcement learning algorithm designed to improve Large Language Model (LLM) reasoning. Traditional methods like GRPO often struggle with "off-policy" data caused by technical mismatches between training and inference engines. OAPL embraces these discrepancies by using a s...

Test-Time Training with KV Binding Is Secretly Linear Attention 27.02.2026

This research paper argues that Test-Time Training (TTT) with key-value binding—previously understood as a way for models to "memorize" data during inference—is actually a form of linear attention. The authors identify a "memorization paradox" where improving the model's internal memory fitting actually degrades task performance, and even reversing the learning process can...

Unified Latents (UL): How to train your latents 26.02.2026

This paper introduces Unified Latents (UL), a novel framework designed by Google DeepMind researchers to improve the efficiency of generative diffusion models. By integrating the encoder, prior, and decoder into a single cohesive system, the method creates a more principled approach to managing latent representations. A key innovation involves using a fixed amount of Gaussian noise during encoding...

Spectral Bellman Method: Unifying RL Representation and Exploration 25.02.2026

This paper introduces the Spectral Bellman Method (SBM), a novel framework designed to enhance value-based reinforcement learning by unifying representation learning and exploration. By leveraging the Inherent Bellman Error (IBE) condition, the authors demonstrate that optimal feature representations are intrinsically linked to the spectral properties of the Bellman operator. This theoretical conn...

Prescriptive Scaling Reveals the Evolution of Language Model Capabilities 24.02.2026

This paper details the development of capability boundaries to predict the downstream performance of language models based on pre-training compute budgets. Researchers move beyond standard scaling laws by using high-quantile pinball loss and monotone splines to track the upper envelope of achievable results rather than average trends. This methodology addresses benchmark saturation, where static t...

Experiential Reinforcement Learning 23.02.2026

**Experiential Reinforcement Learning (ERL)** is a novel training paradigm that enhances how AI agents learn by incorporating a structured **experience-reflection-consolidation loop**. Unlike standard reinforcement learning, which often relies on trial-and-error driven by simple numerical rewards, ERL requires agents to **verbally reflect** on their failures and environment feedback to improve sub...

Слухайте подкаст Best AI papers explained у Replaio

Радіо та подкасти в одному застосунку - безкоштовно й без реєстрації. Встановіть уже сьогодні та не пропустіть запуск

Завантажити з Google Play

Replaio не є видавцем подкастів; назви шоу, обкладинки та аудіо належать їхнім авторам і поширюються через публічні RSS-канали