We argue that language-only models don't learn the physical manifestation of language. We present an empirical investigation of visual-auditory properties of language through a series of tasks, termed H-Test. These tasks highlight a fundamental gap between human linguistic understanding and the sensory-deprived linguistic understanding of LLMs. In support of our hypothesis, 1. deliberate reasoning (Chain-of-Thought), 2. few-shot examples, or 3. stronger LLM from the same model family (LLaMA 2 13B -> LLaMA 2 70B) has no significant effect on H-Test performance. We bring in the philosophical case of Mary, who learns about the world in a sensory-deprived environment as a useful conceptual framework to understand how language-only models learn about the world (Jackson, 1986). Our experiments show that some of the strongest proprietary LLMs stay near random chance baseline accuracy of 50%, highlighting the limitations of linguistic knowledge acquired in the absence of sensory experience. Our code and data are available at .

As language models (LMs) demonstrate their capabilities in various fields, their application to tasks requiring multi-round interactions has become increasingly popular. These tasks usually have complex dynamics, so supervised fine-tuning (SFT) on a limited offline dataset does not yield good performance. However, only a few works attempted to directly train the LMs within interactive decision-making environments. We aim to create an effective approach to fine-tune LMs with online reinforcement learning (RL) in these environments. We propose Reflect-RL, a two-player system to fine-tune an LM using SFT and online RL, where a frozen reflection model (player) assists the policy model (player). To generate data for the warm-up SFT stage, we use negative example generation to enhance the error-correction ability of the reflection model. Furthermore, we designed single-prompt action enumeration and applied curriculum learning to allow the policy model to learn more efficiently. Empirically, we verify that Reflect-RL outperforms SFT and online RL without reflection. Testing results indicate GPT-2 XL 1.56B fine-tuned with Reflect-RL outperforms larger open-source LMs, such as Mistral 7B. The benchmarks, dataset, and code involved in this work are publicly available: https://github.com/zhourunlong/Reflect-RL.

Increasing the number of parameters in large language models (LLMs) usually improves performance in downstream tasks but raises compute and memory costs, making deployment difficult in resource-limited settings. Quantization techniques, which reduce the bits needed for model weights or activations with minimal performance loss, have become popular due to the rise of LLMs. However, most quantization studies use pre-trained LLMs, and the impact of quantization on instruction-tuned LLMs and the relationship between perplexity and benchmark performance of quantized LLMs are not well understood. Evaluation of quantized LLMs is often limited to language modeling and a few classification tasks, leaving their performance on other benchmarks unclear. To address these gaps, we propose a structured evaluation framework consisting of three critical dimensions: (1) knowledge \& capacity, (2) alignment, and (3) efficiency, and conduct extensive experiments across ten diverse benchmarks. Our experimental results indicate that LLMs with 4-bit quantization can retain performance comparable to their non-quantized counterparts, and perplexity can serve as a proxy metric for quantized LLMs on most benchmarks. Furthermore, quantized LLMs with larger parameter scales can outperform smaller LLMs. Despite the memory savings achieved through quantization, it can also slow down the inference speed of LLMs. Consequently, substantial engineering efforts and hardware support are imperative to achieve a balanced optimization of decoding speed and memory consumption in the context of quantized LLMs.

Among recent insights into learning quantum states, online learning and shadow tomography procedures are notable for their ability to accurately predict expectation values even of adaptively chosen observables. In contrast to the state case, quantum process learning tasks with a similarly adaptive nature have received little attention. In this work, we investigate online learning tasks for quantum processes. Whereas online learning is infeasible for general quantum channels, we show that channels of bounded gate complexity as well as Pauli channels can be online learned in the regret and mistake-bounded models of online learning. In fact, we can online learn probabilistic mixtures of any exponentially large set of known channels. We also provide a provably sample-efficient shadow tomography procedure for Pauli channels. Our results extend beyond quantum channels to non-Markovian multi-time processes, with favorable regret and mistake bounds, as well as a shadow tomography procedure. We complement our online learning upper bounds with mistake as well as computational lower bounds. On the technical side, we make use of the multiplicative weights update algorithm, classical adaptive data analysis, and Bell sampling, as well as tools from the theory of quantum combs for multi-time quantum processes. Our work initiates a study of online learning for classes of quantum channels and, more generally, non-Markovian quantum processes. Given the importance of online learning for state shadow tomography, this may serve as a step towards quantum channel variants of adaptive shadow tomography.

We're often told by health experts not to look at our phone just before bedtime as it affects our sleep. But according to a new study, there may not be much scientific basis to this at all. Experts say there's no decent evidence that exposing our eyes to'blue light' from a screen makes it harder to fall asleep. This contradicts official advice from health experts including the NHS, which tells people to avoid using phones an hour before bedtime due to blue light. Instead, the researchers think smartphones are interfering with sleep simply because we can't put them down at night.

Recent Continual Learning (CL) methods have combined pretrained Transformers with prompt tuning, a parameter-efficient fine-tuning (PEFT) technique. We argue that the choice of prompt tuning in prior works was an undefended and unablated decision, which has been uncritically adopted by subsequent research, but warrants further research to understand its implications. In this paper, we conduct this research and find that the choice of prompt tuning as a PEFT method hurts the overall performance of the CL system. To illustrate this, we replace prompt tuning with LoRA in two state-of-the-art continual learning methods: Learning to Prompt and S-Prompts. These variants consistently achieve higher accuracy across a wide range of domain-incremental and class-incremental benchmarks, while being competitive in inference speed. Our work highlights a crucial argument: unexamined choices can hinder progress in the field, and rigorous ablations, such as the PEFT method, are required to drive meaningful adoption of CL techniques in real-world applications.

Robotic affordances, providing information about what actions can be taken in a given situation, can aid robotic manipulation. However, learning about affordances requires expensive large annotated datasets of interactions or demonstrations. In this work, we argue that well-directed interactions with the environment can mitigate this problem and propose an information-based measure to augment the agent's objective and accelerate the affordance discovery process. We provide a theoretical justification of our approach and we empirically validate the approach both in simulation and real-world tasks. Our method, which we dub IDA, enables the efficient discovery of visual affordances for several action primitives, such as grasping, stacking objects, or opening drawers, strongly improving data efficiency in simulation, and it allows us to learn grasping affordances in a small number of interactions, on a real-world setup with a UFACTORY XArm 6 robot arm.

Educational Data Mining has become extremely popular among researchers in last decade. Prior effort in this area was only directed towards prediction of academic performance of a student. Very less number of researches are directed towards predicting employability of a student i.e. prediction of students performance in campus placements at an early stage of enrollment. Furthermore, existing researches on students employability prediction are not universal in approach and is either based upon only one type of course or University/Institute. Henceforth, is not scalable from one context to another. With the necessity of unification, data of professional technical courses namely Bachelor in Engineering/Technology and Masters in Computer Applications students have been collected from 17 states of India. To deal with such a data, a unified predictive model has been developed and applied on 17 states datasets. The research done in this paper proves that model has universal application and can be applied to various states and institutes pan India with different cultural background and course structure. This paper also explores and proves statistically that there is no significant difference in Indian Education System with respect to states as far as prediction of employability of students is concerned. Model provides a generalized solution for student employability prediction in Indian Scenario.

Educational Data Mining (EDM) is a promising field, where data mining is widely used for predicting students performance. One of the most prevalent and recent challenge that higher education faces today is making students skillfully employable. Institutions possess large volume of data; still they are unable to reveal knowledge and guide their students. Data in education is generally very large, multidimensional and unbalanced in nature. Process of extracting knowledge from such data has its own set of problems and is a very complicated task. In this paper, Engineering and MCA (Masters in Computer Applications) students data is collected from various universities and institutes pan India. The dataset is large, unbalanced and multidimensional in nature. A cluster based model is presented in this paper, which, when applied at preprocessing stage helps in parsimonious selection of variables and improves the performance of predictive algorithms. Hence, facilitate in better prediction of Students Employability.

Many people expect robots to move fairly quietly, or make pleasant "beep boop" sounds or jingles similar to what they have observed in videos of robots. Unfortunately, this expectation of quietness does not match reality, as robots make machine sounds, known as 'consequential sounds', as they move and operate. As robots become more prevalent within society, understanding the sounds produced by robots and how these sounds are perceived by people is becoming increasingly important for positive human robot interactions (HRI). This paper investigates how people respond to the consequential sounds of robots, specifically how robots make a participant feel, how much they like the robot, would be distracted by the robot, and a person's desire to colocate with robots. Participants were shown 5 videos of different robots and asked their opinions on the robots and the sounds they made. This was compared with a control condition of completely silent videos. The results in this paper demonstrate with data from 182 participants (858 trials) that consequential sounds produced by robots have a significant negative effect on human perceptions of robots. Firstly there were increased negative 'associated affects' of the participants, such as making them feel more uncomfortable or agitated around the robot. Secondly, the presence of consequential sounds correlated with participants feeling more distracted and less able to focus. Thirdly participants reported being less likely to want to colocate in a shared environment with robots.