In this paper, we explore an important yet previously neglected question: Do context aggregation patterns across Language Models (LMs) share commonalities? While some works have investigated context aggregation or attention weights in LMs, they typically focus on individual models or attention heads, lacking a systematic analysis across multiple LMs to explore their commonalities. In contrast, we focus on the commonalities among LMs, which can deepen our understanding of LMs and even facilitate cross-model knowledge transfer. In this work, we introduce the Order-Level Attention (OLA) derived from the order-wise decomposition of Attention Rollout and reveal that the OLA at the same order across LMs exhibits significant similarities. Furthermore, we discover an implicit mapping between OLA and syntactic knowledge. Based on these two findings, we propose the Transferable OLA Adapter (TOA), a training-free cross-LM adapter transfer method. Specifically, we treat the OLA as a unified syntactic feature representation and train an adapter that takes OLA as input. Due to the similarities in OLA across LMs, the adapter generalizes to unseen LMs without requiring any parameter updates. Extensive experiments demonstrate that TOA's cross-LM generalization effectively enhances the performance of unseen LMs. Code is available at https://github.com/jinglin-liang/OLAS.

Recent advances in large language models, particularly following GPT-4o, have sparked increasing interest in developing omni-modal models capable of understanding more modalities. While some open-source alternatives have emerged, there is still a notable lag behind specialized single-modality models in performance. In this paper, we present Ola, an Omni-modal language model that achieves competitive performance across image, video, and audio understanding compared to specialized counterparts. The core design of Ola lies in its progressive modality alignment strategy that extends the supporting modality of the language model progressively. Our training pipeline begins with the most distinct modalities: image and text, then gradually expands the skill sets of the model using speech data that connects language and audio knowledge, and video data that connects all modalities. The progressive learning pipeline also enables us to maintain a relatively small size of the cross-modal alignment data, making developing omni-modal from existing vision-language models easy and less costly. Moreover, to unlock an advanced interactive experience like GPT-4o, we further design a sentence-wise decoding solution for streaming speech generation. Extensive experiments demonstrate that Ola surpasses existing open omni-modal LLMs across all modalities while achieving highly competitive performance compared to state-of-the-art specialized models of similar sizes. We aim to make Ola a fully open omni-modal understanding solution to advance future research in this emerging field. Model weights, code, and data are open-sourced at https://github.com/Ola-Omni/Ola.

Seasonal climate forecasts are socioeconomically important for managing the impacts of extreme weather events and for planning in sectors like agriculture and energy. Climate predictability on seasonal timescales is tied to boundary effects of the ocean on the atmosphere and coupled interactions in the ocean-atmosphere system. We present the Ocean-linked-atmosphere (Ola) model, a high-resolution (0.25{\deg}) Artificial Intelligence/ Machine Learning (AI/ML) coupled earth-system model which separately models the ocean and atmosphere dynamics using an autoregressive Spherical Fourier Neural Operator architecture, with a view towards enabling fast, accurate, large ensemble forecasts on the seasonal timescale. We find that Ola exhibits learned characteristics of ocean-atmosphere coupled dynamics including tropical oceanic waves with appropriate phase speeds, and an internally generated El Ni\~no/Southern Oscillation (ENSO) having realistic amplitude, geographic structure, and vertical structure within the ocean mixed layer. We present initial evidence of skill in forecasting the ENSO which compares favorably to the SPEAR model of the Geophysical Fluid Dynamics Laboratory.

If you sent me a message on Twitter, email or pigeon post, please give me a few days to dig out of the pile that awaits me. You might recall that I mentioned I was off to do some backpacking and climbing in Grand Teton National Park and then eventually would make it to Yellowstone National Park. Yes, the crowds were real, especially for those who stuck to the traditional schedule of sightseeing between 9 a.m. and 5 p.m. I took the early morning and late evening approach and never encountered the infamous parking lot traffic jams. It's that tactic that allowed me to take a ride in an empty T.E.D.D.Y., the autonomous vehicle that is being piloted in Yellowstone this summer.

Rideshare platform Ola, is using the power of artificial intelligence to help increase safety in the rideshare industry. The new safety-tech feature Guardian will be the first of its kind in New Zealand and is set to roll out in all its locations across the country early next year. The safety technology uses real-time trip information to detect irregular activity such as, a possible crash, an unusually long stop, or an unexpected deviation from the planned route. Ola's 24/7 Safety Response Team is alerted and they contact both riders and drivers to confirm they are safe and offer any assistance they might need. Because Guardian is an intelligent product built using machine learning, it is able to continuously improve its ability to predict risk signals as it keeps collecting data over time, says Ola.

Ola, the popular ride-hailing company this week announced the roll-out of its artificial intelligence-enabled safety feature, Guardian in 17 markets across India and Australia. After running a successful pilot across multiple cities in India and international markets, the Guardian feature is going live in 16 Indian cities as well as Perth in Australia. Ola aims to take Guardian to more cities in the coming quarter. The Guardian feature, developed by Ola as a world-first, uses real-time data from rides to automatically detect irregular trip activity, including prolonged stops and unexpected route deviations. These alerts are flagged off in real-time to Ola's dedicated 24 7 Safety Response Team, who immediately reach out to customers and drivers to confirm if they're safe and offer on-the-call assistance until ride completion.

Ride-hailing major Ola on Monday announced to expand its Artificial Intelligence (AI)-based safety feature called'Guardian' across several cities in the country. 'Guardian' uses real-time data from rides to automatically detect irregular trip activity, including prolonged stops and unexpected route deviations. After running a pilot project, the feature is now live in 16 Indian cities as well as in Perth, Australia and Ola aims to take'Guardian' to more cities in the coming quarter. "We are focused on developing innovations that place customer safety at the heart of platform experience. Guardian brings together the precision of artificial intelligence with the assurance of human intervention, enabling a uniform and safe mobility experience across the markets we operate in," Arun Srinivas, Chief Sales and Marketing Officer, Ola, said in a statement.

Bengaluru-based cab aggregator unicorn Ola has acquired Bengaluru-based artificial intelligence platform Pickup.ai. As part of the deal, the co-founders and employees of Pikup.ai will join the Ola team. When it comes to technology, Ola has been leveraging a lot and has been scaling volumes of success and growth. Its Electric Vehicle arm Ola Electric Mobility became one of the fastest unicorns in India last month. The company too highlighted its commitment and intent of investing in technology especially AI and ML to build futuristic mobility solutions for India and the world.

Indian cab aggregator Ola has announced the acquisition of Bengaluru-based artificial intelligence (AI) startup Pikup.ai on Tuesday. As part of the deal, which value was not disclosed, the team at Pikup.ai including founders Inder Singh and Ritwik Saikia will join Ola. "We are looking forward to joining Ola in its mission to build mobility for a billion people and are very excited about building meaningful technology solutions that have a deep impact on the lives of millions, every single day," said Inder Singh, co-founder of Pikup.ai. The acquisition will fuel Ola's plans to develop next-generation mobility technologies such as autonomous driving, electric and connected mobility. The company said it would do that also by setting up an Advanced Technology Centre in San Francisco.